Contents | Zoom in | Zoom out For navigation instructions please click here Search Issue | Next Page Innovative Technology for Computer Professionals December 2007 A Force of Nature, p. 8 SMS: The Short Message Service, p. 106 Contents | Zoom in | Zoom out __________________ h t t p : / / w w w. c o m p u t e r. o r g Conquering Complexity, p. 111 For navigation instructions please click here Search Issue | Next Page Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F _____________________________ Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Innovative Technology for Computer Professionals Editor in Chief Computing Practices Special Issues Carl K. Chang Rohit Kapur Bill N. Schilit Iowa State University rohit.kapur@synopsys.com ________________ schilit@computer.org ____________ chang@cs.iastate.edu ____________ Bill N. Schilit Kathleen Swigger Michael R. Williams president@computer.org ______________ Web Editor Perspectives Associate Editors in Chief 2007 IEEE Computer Society President Bob Colwell Ron Vetter bob.colwell@comcast.net _______________ vetterr@uncw.edu __________ Research Features Kathleen Swigger University of North Texas kathy@cs.unt.edu __________ Area Editors Column Editors Computer Architectures Steven K. Reinhardt Security Jack Cole Mike Lutz US Army Research Laboratory Edward A. Parrish Databases/Software Michael R. Blaha Broadening Participation in Computing Juan E. Gilbert Embedded Computing Wayne Wolf Software Technologies Mike Hinchey Worcester Polytechnic Institute Modelsoft Consulting Corporation Georgia Institute of Technology Loyola College Maryland Graphics and Multimedia Oliver Bimber Entertainment Computing Michael R. Macedonia Michael C. van Lent How Things Work Alf Weaver Standards John Harauz Reservoir Labs Inc. Bauhaus University Weimar Information and Data Management Naren Ramakrishnan Virginia Tech Multimedia Savitha Srinivasan IBM Almaden Research Center Networking Jonathan Liu University of Florida Software Dan Cooke Texas Tech University Robert B. France Colorado State University Rochester Institute of Technology Ron Vetter University of North Carolina at Wilmington Alf Weaver University of Virginia Jonic Systems Engineering Inc. CS Publications Board Web Technologies Simon S.Y. Shim Jon Rokne (chair), Mike Blaha, Doris Carver, Mark Christensen, David Ebert, Frank Ferrante, Phil Laplante, Dick Price, Don Shafer, Linda Shafer, Steve Tanimoto, Wenping Wang University of Virginia SAP Labs In Our Time David A. Grier Advisory Panel George Washington University University of Virginia IT Systems Perspectives Richard G. Mathieu Thomas Cain James H. Aylor CS Magazine Operations Committee University of Pittsburgh James Madison University Doris L. Carver Invisible Computing Bill N. Schilit The Profession Neville Holmes Louisiana State University Ralph Cavin Semiconductor Research Corp. Ron Hoelzeman University of Pittsburgh Robert E. Filman (chair), David Albonesi, Jean Bacon, Arnold (Jay) Bragg, Carl Chang, Kwang-Ting (Tim) Cheng, Norman Chonacky, Fred Douglis, Hakan Erdogmus, David A. Grier, James Hendler, Carl Landwehr, Sethuraman (Panch) Panchanathan, Maureen Stone, Roy Want University of Tasmania H. Dieter Rombach AG Software Engineering Administrative Staff Editorial Staff Scott Hamilton Lee Garber Senior Acquisitions Editor shamilton@computer.org _____________ Senior News Editor Judith Prow Associate Editor Managing Editor jprow@computer.org ___________ Bob Ward Chris Nelson Bryan Sallis Senior Editor Design and Production Larry Bauer Cover art Dirk Hagner Margo McCall Associate Staff Editor Associate Publisher Dick Price Membership & Circulation Marketing Manager Georgann Carter Business Development Manager Sandy Brown Senior Advertising Coordinator Marian Anderson Publication Coordinator James Sanders Senior Editor Circulation: Computer (ISSN 0018-9162) is published monthly by the IEEE Computer Society. IEEE Headquarters, Three Park Avenue, 17th Floor, New York, NY 100165997; IEEE Computer Society Publications Office, 10662 Los Vaqueros Circle, PO Box 3014, Los Alamitos, CA 90720-1314; voice +1 714 821 8380; fax +1 714 821 4010; IEEE Computer Society Headquarters,1730 Massachusetts Ave. NW, Washington, DC 20036-1903. IEEE Computer Society membership includes $19 for a subscription to Computer magazine. Nonmember subscription rate available upon request. Single-copy prices: members $20.00; nonmembers $99.00. Postmaster: Send undelivered copies and address changes to Computer, IEEE Membership Processing Dept., 445 Hoes Lane, Piscataway, NJ 08855. Periodicals Postage Paid at New York, New York, and at additional mailing offices. Canadian GST #125634188. Canada Post Corporation (Canadian distribution) publications mail agreement number 40013885. Return undeliverable Canadian addresses to PO Box 122, Niagara Falls, ON L2E 6S8 Canada. Printed in USA. Editorial: Unless otherwise stated, bylined articles, as well as product and service descriptions, reflect the author’s or firm’s opinion. Inclusion in Computer does not necessarily constitute endorsement by the IEEE or the Computer Society. All submissions are subject to editing for style, clarity, and space. Published by the IEEE Computer Society Computer 1 December 2007 Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F December 2007, Volume 40, Number 12 Innovative Technology for Computer Professionals IEEE Computer Society: http://computer.org Computer : http://computer.org/computer computer@computer.org ______________ IEEE Computer Society Publications Office: +1 714 821 8380 COMPUTING PRACTICES 24 Examining the Challenges of Scientific Workflows Yolanda Gil, Ewa Deelman, Mark Ellisman, Thomas Fahringer, Geoffrey Fox, Dennis Gannon, Carole Goble, Miron Livny, Luc Moreau, and Jim Myers Workflows have emerged as a paradigm for representing and managing complex distributed computations and are used to accelerate the pace of scientific progress. A recent National Science Foundation workshop brought together domain, computer, and social scientists to discuss requirements of scientific applications and the challenges they present to workflow technologies. COVER FEATURES 33 The Case for Energy-Proportional Computing Luiz André Barroso and Urs Hölzle Energy-proportional designs would enable large energy savings in servers, potentially doubling their efficiency in real-life use. Achieving energy proportionality will require significant improvements in the energy usage profile of every system component, particularly the memory and disk subsystems. 39 Suzanne Rivoire, Mehul A. Shah, Parthasarathy Ranganathan, Christos Kozyrakis, and Justin Meza Power consumption and energy efficiency are important factors in the initial design and day-to-day management of computer systems. Researchers and system designers need benchmarks that characterize energy efficiency to evaluate systems and identify promising new technologies. To predict the effects of new designs and configurations, they also need accurate methods of modeling power consumption. Cover design and artwork by Dirk Hagner ABOUT THIS ISSUE espite advances in power efficiency fueled largely by the mobile computing industry, computer-energy consumption continues to challenge both industry and the global economy. In the US, enterprise energy consumption doubled over the past five years and will continue to do so. And this does not include the energy cost of manufacturing components—it is estimated that Japan’s semiconductor industry will consume 1.7 percent of that country’s energy budget by 2015. The articles in this issue propose strategies for mitigating these costs by designing systems that consume energy in proportion to the amount of work performed, establishing new benchmarks and accurate ways of modeling power consumption, and even recycling older processors over several computing generations. D Computer Models and Metrics to Enable Energy-Efficiency Optimizations 50 The Green500 List: Encouraging Sustainable Supercomputing Wu-chun Feng and Kirk W. Cameron The performance-at-any-cost design mentality ignores supercomputers’ excessive power consumption and need for heat dissipation and will ultimately limit their performance. Without fundamental change in the design of supercomputing systems, the performance advances common over the past two decades won’t continue. 56 Life Cycle Aware Computing: Reusing Silicon Technology John Y. Oliver, Rajeevan Amirtharajah, Venkatesh Akella, Roland Geyer, and Frederic T. Chong Despite the high costs associated with processor manufacturing, the typical chip is used for only a fraction of its expected lifetime. Reusing processors would create a “food chain” of electronic devices that amortizes the energy required to build chips over several computing generations. Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Flagship Publication of the IEEE Computer Society CELEBRATING THE PAST 8 In Our Time A Force of Nature David Alan Grier 11 32 & 16 Years Ago Computer, December 1975 and 1991 Neville Holmes NEWS 13 Industry Trends The Changing World of Outsourcing Neal Leavitt 17 Technology News A New Virtual Private Network for Today’s Mobile World Karen Heyman 20 News Briefs Linda Dailey Paulson MEMBERSHIP NEWS 6 62 64 66 President’s Message Report to Members: Election Results IEEE Computer Society Connection Call and Calendar COLUMNS 103 Security Natural-Language Processing for Intrusion Detection Allen Stone 106 How Things Work NEXT MONTH: Outlook Issue SMS: The Short Message Service Jeff Brown, Bill Shipman, and Ron Vetter 111 Software Technologies Conquering Complexity Gerard J. Holzmann 114 Entertainment Computing Enhancing the User Experience in Mobile Phones S.R. Subramanya and Byung K. Yi 118 Invisible Computing Taking Online Maps Down to Street Level Luc Vincent 124 The Profession Making Computers Do More with Less Simone Santini DEPARTMENTS 4 Article Summaries 23 Computer Society Information 68 IEEE Computer Society Membership Application Computer 72 85 86 102 Annual Index Advertiser/Product Index Career Opportunities Bookshelf COPYRIGHT © 2007 BY THE INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS INC. ALL RIGHTS RESERVED. ABSTRACTING IS PERMITTED WITH CREDIT TO THE SOURCE. LIBRARIES ARE PERMITTED TO PHOTOCOPY BEYOND THE LIMITS OF US COPYRIGHT LAW FOR PRIVATE USE OF PATRONS: (1) THOSE POST-1977 ARTICLES THAT CARRY A CODE AT THE BOTTOM OF THE FIRST PAGE, PROVIDED THE PER-COPY FEE INDICATED IN THE CODE IS PAID THROUGH THE COPYRIGHT CLEARANCE CENTER, 222 ROSEWOOD DR., DANVERS, MA 01923; (2) PRE-1978 ARTICLES WITHOUT FEE. FOR OTHER COPYING, REPRINT, OR REPUBLICATION PERMISSION, WRITE TO COPYRIGHTS AND PERMISSIONS DEPARTMENT, IEEE PUBLICATIONS ADMINISTRATION, 445 HOES LANE, P.O. BOX 1331, PISCATAWAY, NJ 08855-1331. Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ARTICLE SUMMARIES Examining the Challenges of Scientific Workflows pp. 24-32 Yolanda Gil, Ewa Deelman, Mark Ellisman, Thomas Fahringer, Geoffrey Fox, Dennis Gannon, Carole Goble, Miron Livny, Luc Moreau, and Jim Myers W orkflows have recently emerged as a paradigm for representing and managing complex distributed scientific computations, accelerating the pace of scientific progress. Scientific workflows orchestrate the data flow across the individual data transformations and analysis steps, as well as the mechanisms to execute them in a distributed environment. Workflows should thus become firstclass entities in the cyberinfrastructure architecture. Each step in a workflow specifies a process or computation to be executed—a software program or Web service, for instance. The workflow links the steps according to the data flow and dependencies among them. The representation of these computational workflows contains many details required to carry out each analysis step, including the use of specific execution and storage resources in distributed environments. The Case for EnergyProportional Computing pp. 33-37 Luiz André Barroso and Urs Hölzle E nergy management has now become a key issue for servers and data center operations, focusing on the reduction of all energy-related costs, including capital, operating expenses, and environmental impacts. Many energy-saving techniques developed for mobile devices became natural candidates for tackling this new problem space. Although servers clearly provide many parallels to the mobile space, they require additional energy-efficiency innovations. Energy-proportional computers would enable such savings, potentially doubling the efficiency of a typical server. 4 Computer Computer In current servers, the lowest energyefficiency region corresponds to their most common operating mode. Addressing this mismatch will require significant rethinking of components and systems. To that end, energy proportionality should become a primary design goal. Although researchers’ experience in the server space motivates these observations, energy-proportional computing also will significantly benefit other types of computing devices. The Energy-Efficiency Challenge: Optimization Metrics and Models pp. 39-48 Suzanne Rivoire, Mehul A. Shah, Parthasarathy Ranganathan, Christos Kozyrakis, and Justin Meza I n recent years, server and data center power consumption has become a major concern, directly affecting a data center’s electricity costs and requiring the purchase and operation of cooling equipment, which can consume from one-half to one watt for every watt of server power consumption. All these power-related costs can potentially exceed the cost of purchasing hardware. Moreover, the environmental impact of data center power consumption is receiving increasing attention, as is the effect of escalating power densities on the ability to pack machines into a data center. The two major and complementary ways to approach this problem involve building energy efficiency into the initial design of components and systems, and adaptively managing the power consumption of systems or groups of systems in response to changing conditions in the workload or environment. The Green500 List: Encouraging Sustainable Supercomputing pp. 50-55 Wu-chun Feng and Kirk W. Cameron mance per watt has only improved 300-fold and performance per square foot only 65-fold, forcing researchers to design and construct new machine rooms and, in some cases, entirely new buildings. Compute nodes’ exponentially increasing power requirements are a primary driver behind this less efficient use of power and space. Today, several of the most powerful supercomputers on the TOP500 List each require up to 10 megawatts of peak power—enough to sustain a city of 40,000. To inspire more efficient conservation efforts, the HPC community needs a Green500 List to rank supercomputers on speed and power requirements and to supplement the TOP500 List. Life Cycle Aware Computing: Reusing Silicon Technology pp. 56-61 John Y. Oliver, Rajeevan Amirtharajah, Venkatesh Akella, Roland Geyer, and Frederic T. Chong M any consumer electronic devices, from computers to set-top boxes to cell phones, require sophisticated semiconductors such as CPUs and memory chips. The economic and environmental costs of producing these processors for new and continually upgraded devices are enormous. Because the semiconductor manufacturing process uses highly purified silicon, the energy required is quite high—about 41 megajoules (MJ) for a 1.2 cm2 dynamic random access memory (DRAM) die. In terms of environmental impact, 72 grams of toxic chemicals are used to create such a die. Processor reuse can help deal with these increasingly severe economic and environmental costs, but it will require innovative techniques in reconfigurable computing and hardware-software codesign as well as governmental policies that encourage silicon reuse. D espite a 10,000-fold increase since 1992 in supercomputers’ performance when running parallel scientific applications, perfor- Published by the IEEE Computer Society Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F “The technical details and clarity of the articles is beyond anything I’ve seen available elsewhere.” – Sun Microsystems Engineer and IEEE Subscriber From Imagination to Market IEEE Computer Society Digital Library Premier Collection of Computing Periodicals & Conferences Covers the complete spectrum of computing and delivers the highest quality, peer-reviewed content available to users. Over 180,000 top quality computing articles and papers 23 peer-reviewed periodicals Over 170 conference proceedings, with a backfile to 1995 OPAC links for easy cataloging Monthly ‘what’s new’ email notification of new content and services Free Trial! Experience IEEE – request a trial for your company. www.ieee.org/computerlibrary IEEE Information Driving Innovation Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F PRESIDENT’S MESSAGE An Interesting Year Michael R. Williams IEEE Computer Society 2007 President The Society’s 2007 president reflects on a year of challenges and accomplishments. T here is an old phrase that says, “May you live in interesting times.” This saying is often interpreted as a curse because “interesting” can imply a wide variety of situations. I can confidently say that 2007 has been “interesting” in almost every sense of the word. After much debate, we chose Angela Burgess for the job, and she assumed her duties shortly thereafter. Those of you who have met Angela will, I am sure, think that the search committee did a great job in arriving at this decision. I am confident that this will be a major step forward in the Society’s history, and Angela’s abilities will be welcomed for many years to come. A NEW EXECUTIVE DIRECTOR One major event this year was the search for a new executive director of the Computer Society. The search was open, in the sense that there was no obvious candidate. Many extremely well-qualified individuals applied, and reducing the list to manageable proportions was a difficult task for the search committee. The candidates represented diverse areas, including academic, industrial, and nonprofit institutions. Although I have been involved in searching for senior leadership and administrative people in the past, I have seldom seen a more impressive list of individuals than those we chose to interview. After several interviews, we finally met over a holiday weekend to spend both formal and informal time with the top three candidates. They each brought a different set of strengths to the position and each would have made a fine executive director. 6 A BUILDING CRISIS With every good “interesting” development, there is often a counterpart. One of the worst this year was the result of trying to be good citizens within the IEEE. Our staff reorganization (about which, more later) left us with some excess space in our Washington, DC, headquarters building. The building is well over 100 years old and is a heritage-listed structure in the middle of Embassy Row in DC. Another IEEE organization was headquartered in Washington, and it seemed to make eminent sense to agree to a proposal that we share office space in our building. Since this would require some renovations, we called in engineers and architects to advise us—no sense trying to remove a support wall or something equally devastating. When the reports came back, we were surprised to learn that the building infrastruc- ture—electricity, plumbing, heating, and so on—not only could not accommodate the proposed renovations, but was outdated enough to pose potential safety risks. We immediately moved our staff out until we could determine the best course of action to remedy the problems. As I write this message, a second group of engineers is studying the situation, so I can’t give you any definite word on the final outcome. While we would all like instant answers in such situations, doing the job properly takes time. I hope that we can have definitive plans and cost estimates in hand by the time you actually read this. I would like to thank our sister IEEE organization, IEEE-USA, for providing us emergency office accommodations until we sort out this mess. They’ve been extremely helpful and, despite this disruption to their own office functions, have been more than welcoming to our employees. REORGANIZING THE SOCIETY My presidential message at the start of 2007 indicated that this would be a year of decision regarding our organization. I am pleased to say that much of the staff and volunteer reorganization has begun, and I am sure this will result in a more effective organization in the future. Like any major change, the complete plan for reorganizing our Society will take time and will likely be an ongoing effort for several years. Angela Burgess, our new executive director, has been instrumental in implementing the myriad details of such things as rewriting position descriptions and recruiting staff to fill the new vacancies. It is not only the staff that is being reorganized, but also the volunteer side of the organization. While it might be simple to say that combining the Technical Activities Board and the Conferences and Tutorials Board will result in greater synergy in both areas, it is something else to actually plan for a smooth transition, rewrite the governing bylaws, establish new modes of working, and try to foresee potential pitfalls. I would like to Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page express my personal thanks to all the many volunteers who have helped in this and similar endeavors. The list is long, and I will not try to name them all. Of course, such dedicated effort also will be required in 2008 to accomplish the next steps in the plan. All this reorganization effort is necessary for both budget and efficiency reasons. However, it is disruptive and we must not lose sight of the reasons this Society exists in the first place. We have so far managed to keep a good perspective on the situation and not only have kept up such things as our benefits to members but have actually increased them in some areas. For example, in 2008, student members will have a particularly attractive benefit of being able to access free software from Microsoft. I have always tried to remember that making changes is but a step toward providing better services to our constituents. However, common wisdom—particularly that saying about when you’re up to your waist in alligators, it isn’t easy to remember that the objective was to drain the swamp—rings true. AN END AND A BEGINNING As my term as president comes to an end, I look back on this as truly being an “interesting” year. The three items I have touched on in this message are but a fraction of the events and situations that have kept me busy. In my January message in Computer, I said, “I hope that, at the end of this year, we can look back and not only conclude that I did my best but that it was to the benefit of the Society and IEEE as a whole.” I can say that I have done my best, but I will leave it to others to make the rest of that judgment. On 1 January 2008, Ranga Kasturi will take over from me as president, and Susan (Kathy) Land will be the president-elect. Kasturi (as he is usually known) is one of the most thoughtful and capable individuals I have ever met. He will certainly be a president who will take the Society forward to even greater accomplishments. Kathy is also an accomplished and dedicated volunteer, and the two of them will make a good team. I have every confidence that I leave the Society in the best possible hands for 2008 and beyond. A BEMaGS F W ith a Society as large as ours, the personal experiences of our members span the complete spectrum. I have heard from some of you who have gone on to great success in 2007, from others who suffered the ravages of earthquakes and hurricanes, and others who have had less extreme experiences. Whatever your experiences in 2007, I want to wish you the very best possible 2008. It has been my honor to serve as your president in this “interesting” year, and I thank you for the opportunity. I would also be remiss if I did not thank all the dedicated volunteers and staff that made it possible to actually accomplish all that we did in 2007. ■ Michael R. Williams, a professor emeritus of computer science at the University of Calgary, is a historian specializing in the history of computing technology. Contact him at _______________ m.williams@computer. org. __ _______________________________________________ 7 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F IN OUR TIME A Force of Nature David Alan Grier George Washington University Our educational system does little to prepare computer science students for making the transition to the working world. W hen I first met them, Jeff, Jeff, and Will were inseparable. Un pour tous, tous pour un. As far as I could tell, they spent every waking moment in each other’s presence. I could commonly find them at the local coffee shop or huddled together in some corner of a college building. The younger Jeff would be telling an elaborate story to whatever audience was ready to listen. The elder Jeff would be typing on a computer keyboard. Will might be doodling in a notebook or flirting with a passerby or playing garbage can basketball with pages torn from the day’s newspaper. Among colleagues, I referred to them as the Three Musketeers, as they seemed to embody the confidence of the great Dumas heroes. They were masters of technology and believed that their mastery exempted them from the rules of ordinary society. The younger Jeff, for example, believed that he was not governed by the law of time. When given a task, he would ignore it until the deadline was bearing down on him. Then in an explosion of programming energy, he would pound perfect code into his machine. The elder Jeff was convinced that specifications were written for other 8 people, individuals with weaker morals or limited visions. He wrote code that was far grander than the project required. It would meet the demands of the moment, but it also would spiral outward to handle other tasks as well. You might find a game embedded in his programs or a trick algorithm that had nothing to do with the project or a generalization that would handle the problem for all time to come. Will was comfortable with deadlines and would read specifications but he lived in a non-Euclidian world. He shunned conventional algorithms and obvious solutions. His code appeared inverted, dissecting the final answer in search of the original causes. It was nearly impossible to read, but it worked well and generally ran faster than more straightforward solutions. DISRUPTION The unity of the Three Musketeers was nearly destroyed when Alana came into their circle. She was a force of nature and every bit the intellectual equal of the three boys. She took possession of their group as if it were her private domain. Within a few weeks, she had them following her schedule, meeting at her favorite places, and doing the things that she most liked to do. She even got them to dress more stylishly, or at least put on cleaner clothes. Alana could see the solution of a problem faster than her compatriots, and she knew how to divide the work with others. For a time, I regularly saw the four of them in the lounge, laughing and boasting as they worked on some class project. One of their number, usually a Jeff, would be typing into a computer while the others discussed what task should be done next. Paper wads would be scattered around a wastebasket. A clutch of pencils would be neatly balanced into a pyramid. It was not inevitable that Alana should destabilize the group, but that is what eventually happened. Nothing had prepared the boys for a woman who had mastered both the technical details of multiprocessor coding and the advanced techniques of eye makeup. For reasons good or ill, Alana was able to paint her face in a way that made the souls of ordinary men melt into simmering puddles of sweat. Steadily, the group began to dissolve. The end was marked with little, gentle acts of kindness that were twisted into angry, malicious incidents by the green-eyed monster of jealousy. Soon Jeff was not speaking to Jeff, Will was incensed with the Elder, the Younger had temporarily decamped for places unknown, and Alana was looking for a more congenial group of colleagues. Eventually, the four were able to recover the remnants of their friendship and rebuild a working relationship, but they never completely recovered their old camaraderie. Shortly, they moved to new jobs and new worlds, where they faced not only the pull of the opposite sex but also had to deal with the equally potent and seductive powers of finance and money. MOVING ON The younger Jeff was the first to leave. He packed his birthright and followed the western winds, determined to conquer the world. With a Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page few friends, he built an Internet radio station, one of the first of the genre. They rented space in a warehouse, bought a large server, and connected it to the Internet. They found some software to pull songs off their collection of CDs and wrote a system that would stream music across the network while displaying ads on a computer screen. They christened their creation “intergalactic-radio-usa.net.” __________________ For a year or two, the station occupied a quirky corner of the Net. It was one of the few places in those days before MP3 downloads where you could listen to music over the Internet. Few individuals had a computer that reproduced the sounds faithfully, but enough people listened to make the business profitable. One day, when he arrived at work, Jeff was met at the door by a man with a dark overcoat, a stern demeanor, and a letter from one of the music publishing organizations, BMI or ASCAP. The letter noted that intergalactic-radio-usa had not been paying royalties on the music it broadcast, and it demanded satisfaction. A date was set. Seconds were selected. Discussions were held. Before the final confrontation, the station agreed to a payment schedule and returned to the business of broadcasting music. Under the new regime, the station had to double or triple its income in a short space of time. This pushed Jeff away from the technical work of the station. His partners had grown anxious over his dramatic programming techniques. They wanted steady progress toward their goals, not weeks of inaction followed by days of intense coding. They told him to get a suit, build a list of clients, and start selling advertising. Jeff was not a truly successful salesman, but he also was not a failure. His work allowed him to talk with new people, an activity he loved, but it did not give him the feeling of mastery that he had enjoyed as a programmer. “It’s all governed by the budget,” he told me when he started to look for a new job. “Everything is controlled by the budget.” AN EVOLVING BUSINESS The elder Jeff left shortly after his namesake. He and some friends moved into an old house in a reviving part of the city, a living arrangement that can best be described as an Internet commune. They shared expenses and housekeeping duties and looked for ways to make money with their computing skills. Slowly, they evolved into a Web design and hosting company. They created a Web page for one business and then one for another and finally admitted that they had a nice, steady line of work. Overcoming the pressures and demands of the commercial world requires qualities learned over a lifetime. As the outlines of their company became clearer, Jeff decided that they were really a research and development laboratory that supported itself with small jobs. He took a bedroom on an upper floor and began working on a program that he called “the ultimate Web system.” Like many of the programs the elder Jeff produced, the ultimate Web system was a clever idea. It is best considered an early content management system, a way to allow ordinary users to post information without working with a programmer. As good as it was, the ultimate Web system never became a profitable product. Jeff had to abandon it as he and his partners began to realize that their company needed stronger leadership than the collective anarchy of a commune. They needed to coordinate the work of designers, programmers, salespeople, and accountants. As the strongest personality of the group, Jeff slowly moved into the role of president. As he did, the company became a more conventional organization. Couples married and moved into their own homes. The large house A BEMaGS F uptown ceased to be a residence and became only an office. Long after Jeff had begun purchasing Web software, he continued to claim that he was a software developer. “I’ll get back to it some day,” he would say. “It will be a great product.” MAKING CHOICES Will was the last to leave. He started a company that installed computers for law firms. Our city hosts a substantial number of lawyers, so his long-term prospects were good. I once saw Will on the street, pushing a cart of monitors and network cables. He looked active and happy. Things were going well, he said. He had plenty of work but still had enough time to do a little programming on the side. We shook hands, promised to keep in touch, and agreed to meet for dinner on some distant day. That dinner went unclaimed for five years. It might have never been held had not I learned, through one of the Jeffs, that Will had prospered as a programmer and now owned a large specialty software firm. I scanned his Web page and was pleased with what I saw. “Software in the service of good,” it read. “Our motto is people before profit. Do unto others as you would have them do unto you.” I called his office, was connected to “President Will,” and got a quick summary of his career. He had started creating programs for disabled users and had found a tremendous market for his work. After a brief discussion, we agreed to a nice dinner downtown, with spouses and well-trained waiters and the gentle ambience of success. We spent most of the evening talking about personal things—families, children, and houses. Only at the end of the evening did we turn to work. “How is the business going?” I asked. “Well,” he said, but the corners of his lips pursed. I looked at him a moment. “Everything okay?” I queried. He exchanged a glance with his wife and turned to back to me. “Extremely 9 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F IN OUR TIME well. We have more business than we can handle.” Again, I paused. “Starting to draw competition?” He smiled and relaxed for a brief moment. “No,” he said. I sensed that something was happening, so I took a guess. “A suitor sniffing around?” He looked chagrined and shook his head. “Yeah.” “A company with three letters in its name?” “Yup,” he said. “It would be a lot of money,” I noted. “But then it wouldn’t be my firm,” he said. After a pause, he added, “And if I don’t sell, the purchaser might try to put me out of business.” We moved to another subject, as Will was clearly not ready to talk any more about the potential sale of his company. It was months later that I learned that he had sold the company and had decided to leave the technology industry. The news came in a letter that asked me to write a recommendation for a young man who wanted to become “Reverend Will.” Almost every technical field feels the constant pull of business demands. “Engineering is a scientific profession,” wrote the historian Edwin Layton, “yet the test of the engineer’s work lies not in the laboratory but in the marketplace.” By training, most engineers want to judge their work by technical standards, but few have that opportunity. “Engineering is intimately related to fundamental choices of policy made by organizations employing engineers,” notes Layton. LESSONS LEARNED The experiences of Jeff, Jeff, and Will have been repeated by three decades of computer science students. They spend four years studying languages, data structures, and algorithms and pondering that grand question, “What can be automated?” Then they leave that world and move to one in which profit is king, deadlines are queens, and finance is a knave that keeps order. Our educational system does little to prepare them for this transition. An early report reduced the issue to a pair of sentences. “A large portion of the job market involves work in business-oriented computer fields,” the report noted before making the obvious recommendation. “As a result, in those cases where there is a business school or related department, it would be most appropriate to take courses in which one could learn the technology and techniques appropriate to this field.” Of course, one or two courses can’t really prepare an individual for the pressures and demands of the commercial world. Overcoming pressures requires qualities that are learned over a lifetime. Individuals need poise, character, grace, a sense of right and wrong, an ability to find a way through a confused landscape. Often professional organizations, including many beyond the field of computer science, have reduced such qualities to the concept of skills that can be taught in training sessions: communications, teamwork, self-confidence. In fact, these skills are better imparted by the experiences of life, by The IEEE Computer Society publishes over 150 conference publications a year. For a preview of the latest papers in your field, visit www.computer.org/publications/ 10 learning that your roommate is writing and releasing virus code, that you have missed a deadline and will not be paid for your work, that a member of your study group is passing your work as his own. “It is doubly important,” wrote Charles Babbage, “that the man of science should mix with the world.” In fact, most computer scientists have little choice but to mix with the world, as the world provides the discipline with problems, ideas, and capital. It is therefore doubly important to know how the ideas of computer science interact with the world of money. O ne of the Three Musketeers, safely out of earshot of his wife, once asked me what had become of Alana. I was not in close contact with her. However, I knew that her life had been shaped by the same forces that had influenced the careers of her three comrades, although, of course, her story had been complicated by the issues that women must face. She had moved to Florida and built a business during the great Internet bubble. She had taken some time to start a family, perhaps when that bubble had burst in 2001, and was now president of her own firm. I didn’t believe that she had done technical work for years. “You’ve not kept in touch,” I said more as a statement than a question. “No,” he said. I saw a story in his face, but that story might reflect more of my observations than of his experience. It told of a lost love, perhaps; a lesson learned; a recognition that he and his college friends were all moving into that vast land of the mid-career knowing a little something about how to deal with the forces of nature. ■ David Alan Grier is the editor in chief, IEEE Annals of the History of Computing, and the author of When Computers Were Human (Princeton University Press, 2005). Grier is associate dean of International Affairs at George Washington University. Contact him at ____ grier@ gwu.edu. ______ Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F 1975•1991• 1975•1991•1975•1991•1975•1991 32 & 16 YEARS AGO DECEMBER 1975 COMPUTER EDUCATION (p. 27). “From the educator’s point of view, perhaps no problem is so apparent as that of overcoming the dichotomy between computer science and computer engineering. the task of developing curricula that harmoniously integrate those two components of computing has been reminiscent of the battles of great prehistoric beasts in the tar pits: the fiercer and more passionate the struggle, the sooner the combatants are ensnared in the tar. So far at any rate, the student has had to choose between computer science or computer engineering—and of course both he and his employer have had to pay the price in terms of increased training requirements and delayed effectiveness.” COMPUTING CURRICULA (p. 29). “It is the conclusion of [the IEEE Computer Society Model Curricula Subcommittee] that these four areas will provide a first entry to industry-level instruction in computer science and engineering: “Digital/Processor Logic Instruction towards an understanding of digital logic devices and their interconnection to provide processing functions. … “Computer Organization Instruction towards an understanding of the interaction between computer hardware and software, and the interconnection of system components. “Operating Systems and Software Engineering “Theory of Computing Instruction towards an understanding of the formal aspects of computing concepts, to include discrete structure, automata and formal languages, and analysis of algorithms.” US SURVEY (p. 40). “The growth of computer science and computer engineering, at least in EE departments, has slowed but still continues. Throughout the survey, growth since 1972 has been slower than growth before 1972. EE departments offer only 2% more CS or CE options than they did two years ago. They have slightly more faculty in the computer area. They do offer more computer courses, particularly minicomputer and microcomputer courses. The growth of the latter courses has been very rapid with 37% of EE departments reporting one or more courses on microprocessors in the 19741975 school year.” LARGE-SCALE COMPUTERS (p. 82). “Amdahl Corporation has delivered three of its $3.7 to $6.0 million 470V/6 large-scale computers and expects to deliver three more before the end of the year. In all installations to date (Texas A&M University, Columbia University, and the University of Michigan), the computer replaced one or more IBM systems. “The company states that the 470V/6 can be substituted for an IBM 370 or 360 being used to run any set of programs and using any peripheral mix. No more changes are required than would be to change from one model of the 370 to another.” COMPUTER MEDIA SERVICING (p. 84). “A new approach in the servicing of computer centers will be pursued jointly in the Los Angeles area by Memorex and Datavan Corporations. “Specially-outfitted and staffed Datavan mobile units will travel to customer computer installations where they will re-ink ribbons, clean and recertify tapes, and clean disk packs. They will be stocked with Memorex’s computer media products consisting of computer tape, disk packs and cartridges, data modules, and flexible disks.” IMPACT PRINTER (p. 86). “Documation, Inc. has announced the availability of the new DOC 2250 highspeed, impact line printer, capable of printing 2250 single-spaced lines per minute using a 48 graphic character set.” “The printer is a free-standing unit containing its own power supply and control logic. The integrated controller is a Documation-developed microprocessor. The controller communicates through its interface with the host system, decodes all commands, controls the printer hardware, and reports various errors and status.” PANAMA CANAL (p. 88). “About 40 ships per day pass through the Panama Canal and that number is expected to increase significantly in coming years. Because of the projected upswing, the Panama Canal Company performed studies to develop requirements and specifications for a new Marine Traffic Control System (MTCS). “The MTCS, implementing 25 General Electric TermiNet 300 send-receive printers, is a computer-based system designed to assist in collecting, assimilating, displaying, and disseminating schedules and related data used to coordinate and control transit operations.” POST OFFICE LABELS (p. 88). “The United States Postal Service will introduce a labeling system next year which will increase the efficiency of transporting mail between the post office’s 40,000 stations around the country. “The system, being developed by the Electronics and Space Division of Emerson Electric Company, is essentially a computerized printing system which has a capability of producing a yearly total of eight-billion labels and slips used to attach to mail bags and bundles of letters with a common destination.” “With the new system, the post office’s computerized printing plant for labels and slips will produce a supply for every post office based on its own special needs. The necessary data will be stored on magnetic tape and used to produce two-week supplies for each order. Changes to the order will be easily assimilated on the tape, reflecting each office’s latest routing needs.” 11 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page DECEMBER 1991 PRESIDENT’S REPORT (p. 4). “… At its November 1 meeting, the Board of Governors approved an agreement with ACM to form the Federation on Computing in the United States (FOCUS), the successor to the American Federation of Information Processing Societies (AFIPS). That agreement establishes an organization of technical societies to represent US computing interests in the International Federation for Information Processing (IFIP).” “We must develop ways to disseminate our transactions, magazines, conferences, and other material electronically. … I have become convinced that the effort required is beyond what we can expect from a few volunteers. For that reason, we asked and obtained board approval for a staff position dedicated to this effort.” DATABASE INTEGRATION (p. 9). “Many believe that standards development will resolve problems inherent in integrating heterogeneous databases. The idea is to develop systems that use the same standard model, language and techniques to facilitate concurrent access to databases, recovery from failures, and data administration functions. This is easier said than done. Agreement on standards has proven to be one of the most difficult problems in the industry. Most vendors and end users have already invested in separate solutions for their problems. Getting them to agree on a common way of handling their data is challenging.” ADDRESSING HETEROGENEITY (p. 17). “Our approach to addressing schematic heterogeneity is to define views on the schemas of more than one component database and to formulate queries against the views. The view definition can specify how to homogenize the schematic heterogeneity in CDBS [Component DataBase System] views. Our approach to data heterogeneity is twofold: First, we allow the MDBS [MultiDataBase System] query processor to issue warnings when it detects wrong-data conflicts in query results. Second, we allow the MDBS users and/or database administrator to prepare and register lookup tables in the database so that the MDBS query processor can match different representations of the same data.” MULTIDATABASE TRANSACTIONS (p. 28). “A transaction is a program unit that accesses and updates the data in a database. An everyday example of a transaction is the transfer of money between bank accounts. The debiting of one account and the crediting of another are each separate actions, yet the combination of these actions is viewed by the ‘user’ as one. The notion of combining several actions into a single logical unit is central to many of the properties associated with transactions.” “Transaction processing in a distributed environment is complex because the actions that compose a transaction can occur at several sites. Either all these actions 12 A BEMaGS F should succeed, or none should. Thus, an important aspect of transaction processing for a distributed system is reaching agreement among sites. The most widely used solution to this agreement problem is the twophase commit (2PC) protocol.” INTERDATABASE CONSISTENCY (p. 46). “In most applications, the mutual consistency requirements among multiple databases are either ignored or the consistency of data is maintained by the application programs that perform related updates to all relevant databases. However, this approach has several disadvantages. First, it relies on the application programmers to maintain mutual consistency of data, which is not acceptable if the programmers have incomplete knowledge of the integrity constraints to be enforced. … Since integrity requirements are specified within an application, they are not written in a declarative way. If we need to identify these requirements, we must extract them from the code, which is a tedious and error-prone task.” AUTONOMOUS TRANSACTIONS (pp. 71-72). “The evolution of classic TP [Transaction Processing] to autonomous TP has just begun. The issues are not yet clearly drawn, but they undoubtedly affect the way TP systems are designed and administrated Transaction execution is affected as well if users require independent TP operations during network partitions or communication failures. We believe that asynchronous TP provides a suitable mechanism to support execution autonomy, because temporary inconsistency can be tolerated and database consistency restored after the failure is repaired. Divergence control methods must be devised that give systems the flexibility needed to evolve from classic TP to autonomous TP.” GENETIC ALGORITHMS (p. 93). “NovaCast of Sweden has launched the C Darwin II general-purpose tool for solving optimization problems. The shell can reproduce with stepwise, leaping changes and an adaptive selection process that accumulates small, successive, and favorable variations from one generation to another. “The program solves product design and planning, machine scheduling, composition, and functional optimization problems. Users define how they want the solution to be presented and describe the environment in terms of its conditions, restrictions, and cost factors. The program evaluates a generation of conceivable solutions in parallel.” PDFs of the articles and departments of the December 1991 issue of Computer are available through the Society’s Web site, www.computer.org/computer. Editor: Neville Holmes; __________________ neville.holmes@utas.edu.au Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F INDUSTRY TRENDS The Changing World of Outsourcing Neal Leavitt O utsourcing—a practice once considered controversial—has become widespread, not only with technology companies but also with the IT departments of firms in other industries. The volume of tech offshore outsourcing—in which companies in economically advanced countries send work to businesses in developing nations—has increased since the approach became popular during the economic boom of the mid1990s. During that time, the nature of the practice stayed largely the same. Now, though, this appears to be changing. For example, companies that handle outsourcing are beginning to consolidate, creating larger providers offering a broader range of services. At the same time, niche providers are emerging. In addition, countries such as China are beginning to compete with India, the longtime outsourcing-services leader. Once, large companies did most tech outsourcing. Now, smaller and mid-size companies are beginning to outsource work. Also, companies primarily used to outsource large projects, such as basic application development or call-center operations. Now, as outsourcing becomes more widespread, businesses are starting to contract out smaller SHORT-TERM TRENDS projects—including complex scientific and R&D projects—to more and more providers. Various technical and marketplace developments have driven and enabled these changes. And industry observers expect more changes in the long run. INSIDE OUTSOURCING Technology-related outsourcing began in the early 1980s and grew rapidly in the mid-1990s. The driving forces included the expanding tech economy, increased pressure on IT departments to do more with their resources, the increasing complexity of managing IT and keeping up with rapidly advancing technologies, and the difficulty in finding IT workers in all skill areas, noted David Parker, vice president of IBM Global Technology Services’ Strategic Outsourcing operations. Since then, companies have outsourced more than just IT functions. For example, they have used the approach to make their production, customer service, and other processes Published by the IEEE Computer Society Computer efficient and inexpensive by farming them out to businesses that have the necessary expertise and can perform them less expensively, Parker noted. This work includes help desk, data center, and network management operations; database administration; and server management. There are outsourcing providers in both economically advanced and developing countries. The latter are able to offer lower costs because workers there receive lower wages. Offshore outsourcing has become controversial, especially in the US, where critics say it is a way for domestic companies to save money by taking jobs from local workers and moving them overseas. Technology has helped change the face of outsourcing. For example, improvements in telecommunications and Internet-based technologies such as videoconferencing, instant messaging, and Internet telephony make communication faster and more widely available for outsourcing providers and their clients, noted Alex Golod, vice president of business development for Intetics, a global software development company. Legal issues are also important. Some businesses in developed nations are outsourcing work to overseas branches of domestic companies to have more legal recourse if problems occur, said Ashish Gupta, CEO of Evalueserve Business Research, a market-analysis firm. Consolidation Because more companies are farming out a greater variety of projects, there has been a proliferation of outsourcing suppliers in recent years to meet the demand. Meanwhile, as outsourcing companies have grown, they have looked for ways to reinvest profits, expand, and acquire new capabilities, noted Gupta. For these reasons, outsourcing is experiencing a number of mergers and acquisitions. 13 December 2007 Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F INDUSTRY TRENDS For instance, the US-based EPAM Systems became the largest software-engineering-services provider in Central and Eastern Europe by acquiring Vested Development, a Russian software-developmentservices firm. India’s Wipro acquired Infocrossing, a US infrastructuremanagement services provider, for $600 million. Wipro has since opened several software-development and IT-services offices in the US. Among other recent outsourcing deals, IBM purchased Daksh, Computer Sciences Corp. bought Covansys, and Electronic Data Systems acquired MphasiS. According to Intetics’ Golod, consolidation is a way for small and midsize vendors to survive and grow while gaining efficiencies in marketing and operations, and for larger companies to acquire competitors and firms with niche skills. Beyond India India has been the leader in outsourcing services since the mid1990s, mostly because it has a large, educated, English-speaking tech workforce; low salaries; and a technology sector that has pursued the work for many years. Market-research firm Gartner warns that a shortage of tech workers and rising wages could erode as much as 45 percent of India’s market share by next year. In addition, Indian outsourcers can’t handle the high volume of available projects. According to Intetics’ Golod, competition is coming largely from places such as China, the Philippines, Eastern Europe, Latin America, and even developed countries like Ireland and Israel, as discussed in the sidebar “Developing Countries Join the Outsourcing Marketplace.” Market-research firm IDC predicts that China will overtake India by 2011. Emerging low-wage countries that might also pull business from India over the next few years include Egypt, Malaysia, Pakistan, and Thailand. Small-scale and niche outsourcing Traditionally, outsourcing has entailed big companies and large, long-term projects such as major application development or the operations of entire departments. Bigger companies have been more willing to pay for outsourcing than smaller ones, have had more tasks they could offload, and could provide larger contracts than smaller businesses. Now, though, more companies are competing with large outsourcing providers. This includes outsourcing providers targeting smaller businesses and smaller, shorter-term jobs—including minor testing and business-analysis projects—either to Developing Countries Join the Outsourcing Marketplace India has been the leading offshore technologyoutsourcing supplier for a decade. The country has leveraged several advantages, including low salaries and a big, English-speaking workforce with college degrees in technology fields. However, there have been growing opportunities for other developing countries to enter the outsourcing-services market. For example, there is more outsourcing work than India can handle. In addition, wages in India are rising, and companies in other countries are actively pursuing and promoting their services. China This country attracts outsourcing projects largely in areas such as low-end, PC-based application development, quality-assurance testing, system integration, data processing, and product development. India is even outsourcing work to China. Market research firm IDC’s Global Delivery Index— which ranks locations according to criteria such as available skills, political risk, and labor costs—said Chinese cities have made significant investments in infrastructure, English-language instruction, and Internetconnection availability. The country has undergone a massive telecommunica- 14 tions expansion as a result of national economic policy. Also, China is producing 400,000 college graduates in technology fields annually, said Kenneth Wong, managing partner at SmithWong Associates, a China-focused US consulting firm. “Language issues are no longer the major handicap to China-based outsourcing,” noted Wong. “Many IT personnel in China today are US-educated. The Chinese government knows it has a way to go in reaching English-proficiency parity with India, but the signs are encouraging.” Paul Schmidt, a partner with outsourcing consultancy TPI, added that demand for the country’s outsourcing services is driven largely by multinational corporations looking for access to China’s domestic market and a presence in the rest of the Asia-Pacific region. Russia Russia’s principal outsourcing competencies include Internet programming, Web design, and Web-server and Web-database application development. The country has an educated, experienced labor force, with the world’s third-largest pool of engineers and scientists per capita. English competency is good for mid- to higher-level managers and acceptable for Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page capture a market niche or to expand their customer base. This work can entail either parts of larger jobs or small, individual projects. To remain competitive, the larger providers are also looking at the smaller end of the outsourcing market, in part by standardizing their offerings to make them less expensive, noted IBM’s Parker. Outsourcing consultancy TPI says the number of commercial outsourcing contracts of more than $50 million, especially in the US, has declined. According to TPI, many Indian providers are seeking smaller jobs than in the past and thus are expected to increase revenue from North American companies by 37 percent during the next two years, even though the region is outsourcing less work. Niche outsourcing providers are well placed for the smaller contracts—often for more focused projects—that are becoming more popular, said Peter Allen, TPI partner and managing director of marketing development. The growth of niche outsourcing is contributing to multisourcing, in which companies farm out different parts of a project to multiple specialty service providers, said Brian C. Daly, public relations director for Unisys’ outsourcing business. “This makes it possible to choose best-of-breed outsourcing firms to handle different tasks and to optimize costs,” noted Intetics’ Golod. Outsourcing new and more complex tasks As outsourcers have gained more skills, companies are beginning to farm out more difficult technology tasks. Banks and other businesses, for instance, are moving well beyond outsourcing low-level applicationmaintenance work and are increasingly relying on offshore providers for help with full-system projects. F Cost not the only factor Companies are farming out work not only to save money but also to gain long-term access to outsourcing firms’ talent, as well as their innovative, creative, and advanced approaches, said Unisys’ Daly. Many companies find recruiting to be a tedious and expensive advantage of time-zone proximity to the US. This makes Latin America ideal for time-sensitive projects and work where the outsourcer must communicate quickly and regularly with the outsourcing company. Eastern Europe The Philippines Countries such as Belarus, Bulgaria, the Czech Republic, Hungary, Poland, Romania, and Ukraine specialize largely in application development, particularly for complex scientific projects or commercial products. The region has a solid educational base, producing qualified scientists and engineers, explained Golod. Also, he said, the workers in these companies prefer complex, challenging projects to simple coding. In addition, larger clients are interested in sending work to Eastern Europe because they want to diversify their outsourcing across geographic regions. Eastern Europe gets a lot of projects from Western Europe and the UK, which are in nearby time zones, as well as the US, he said. This nation has a large English-speaking population and is carving out a niche for call-center operations. Evalueserve Business Research, a market-analysis firm, reports that favorable factors include a 94 percent literacy rate, a high-quality telecommunications infrastructure, familiarity with Western corporate culture, and government initiatives such as exemptions from license fees and export taxes that have stimulated outsourcing growth. Also, wages are low. Countries in this region—particularly Argentina, Brazil, and Mexico—specialize in outsourcing projects such as application development. They have the BEMaGS These include providing online banking capabilities, customizing enterprise-resource-planning systems, or conducting statistical and actuarial projects for insurance firms. Outsourcing the modernization of legacy applications is another area with significant potential. This would help, for example, companies that no longer have the in-house expertise to work with older applications written in languages such as Cobol and that don’t want to spend the time and money necessary to rewrite them in other languages. developers who communicate directly with clients, noted Alex Golod, vice president of business development for Intetics, a global software-development company. Latin America A Developed countries Ireland has a favorable IT services infrastructure and is strong in software development and testing. Israeli outsourcing providers specialize in commercial software development, particularly security and antivirus products. Although both nations have higher labor costs than developing countries, they continue to attract business because of their well-educated workforces and stable governments. 15 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F INDUSTRY TRENDS process, with choices often limited by local manpower resources, added Rob Enderle, president and principal analyst with the Enderle Group, a market-research firm. Global outsourcing firms have access to extensive pools of talent and well-established recruiting procedures. I ndustry observers expect offshore outsourcing to continue growing. For example, Forrester Research, a market-analysis firm, estimates that software outsourcing will account for 28 percent of IT budgets in the US and Europe by 2009, up from 20 percent in 2006. Forrester also predicts that the number of overseas software developers working on projects for firms in developed countries will rise from 360,000 this year to about 1 million by 2009. Companies will increasingly outsource telecommunications- and Internet-based development because the technologies in these areas are based on international standards and are thus easy for offshore vendors to work on. Security is a big concern for many companies, with the growing use of mobile technology and the increasing complexity of threats. More companies will thus outsource shortterm and ongoing security efforts, as keeping up with the work themselves will require too much time and money. However, there may be a limit on growth prospects, Golod said. For example, he noted, international political and economic problems could reduce the amount of outsourced work. According to Enderle, outsourcers might face problems coordinating their work with customers as they Windows Kernel Source and Curriculum Materials for Academic Teaching and Research. The Windows® Academic Program from Microsoft® provides the materials you need to integrate Windows kernel technology into the teaching and research of operating systems. The program includes: • Windows Research Kernel (WRK): Sources to build and experiment with a fully-functional version of the Windows kernel for x86 and x64 platforms, as well as the original design documents for Windows NT. • Curriculum Resource Kit (CRK): PowerPoint® slides presenting the details of the design and implementation of the Windows kernel, following the ACM/IEEE-CS OS Body of Knowledge, and including labs, exercises, quiz questions, and links to the relevant sources. take on a greater variety and number of clients and projects. And, said Renga Rajan, technical director for Fast Pvt. Ltd. (www. ____ fastindia.com), an Indian outsourc__________ ing supplier, providers will face challenges keeping their prices down because of rising salaries for skilled workers in developing countries. ■ Neal Leavitt is president of Leavitt Communications (www.leavcom.com), a Fallbrook, California-based international marketing communications company with affiliate offices in Brazil, France, Germany, Hong Kong, India, and the UK. He writes frequently on technology topics and can be reached at _____________ neal@leavcom.com. Editor: Lee Garber, Computer, l.garber@computer.org ______________ Looking for an “Aha” idea? Find it in CSDL Computer Society Digital Library 200,000+ articles and papers Per article: • ProjectOZ: An OS project environment based on the SPACE kernel-less OS project at UC Santa Barbara, allowing students to develop OS kernel projects in user-mode. $9US (members) These materials are available at no cost, but only for non-commercial use by universities. $19US (nonmembers) For more information, visit www.microsoft.com/WindowsAcademic or e-mail _____________ compsci@microsoft.com. 16 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F TECHNOLOGY NEWS A New Virtual Private Network for Today’s Mobile World Karen Heyman V irtual private networks were a critical technology for turning the Internet into an important business tool. Today’s VPNs establish secure connections between a remote user and a corporate or other network via the encryption of packets sent through the Internet, rather than an expensive private network. However, they traditionally have linked only a relatively few nodes that a company’s IT department controls and configures. This is not adequate for the many organizations that now must let managers, employees, partners, suppliers, consultants, e-commerce customers, and others access networks from their own PCs, laptops, publicly available computers like those at airport kiosks, and even mobile devices, many not controlled by the organization. VPNs based on Internet Protocol security (IPsec) technology were not designed for and are not well-suited for such uses. Instead of restricting remote users who should not have access to many parts of a company’s network, explained Graham Titterington, principal analyst with market-research firm Ovum, “IPsec [generally] connects users into a network and gives the same sort of access they would have if they were physically on the LAN.” Organizations are thus increasingly adopting VPNs based on Secure Sockets Layer technology from vendors such as Aventail, Cisco Systems, F5 Networks, Juniper Networks, and Nortel Networks. SSL VPNs enable relatively easy deployment, added Chris Silva, an analyst at Forrester Research, a market-research firm. A company can install the VPN at its headquarters and push any necessary software to users, who then access the network via their browsers, he explained. Organizations thus don’t have to manage, fix, update, or buy licenses for multiple clients, yielding lower costs, less maintenance and support, and greater simplicity than IPsec VPNs, Silva said. “From a remote-access perspective, IPsec is turning into a legacy technology,” said Rich Campagna, Juniper’s SSL VPN product manager. Nonetheless, IPsec VPNs are still preferable for some uses, such as linking a remote, company-controlled node, perhaps in a branch office, with the corporate network. Both VPN flavors are likely to continue to flourish, with the choice Published by the IEEE Computer Society Computer coming down to a corporation’s physical setup and access needs. However, Silva said, SSL VPNs might eventually have the edge as the world goes more mobile. Meanwhile, SSL VPNs still face some challenges to widespread adoption. VPN BACKGROUND An early attempt to create a VPN over the Internet used multiprotocol label switching, which adds labels to packets to designate their network path. In essence, all packets in a data set travel through designated tunnels to their destinations. However, MPLS VPNs don’t encrypt data. IPsec and SSL VPNs, on the other hand, use encrypted packets with cryptographic keys exchanged between sender and receiver over the public Internet. Once encrypted, the data can take any route over the Internet to reach its final destination. There is no dedicated pathway. US Defense Department contractors began using this technique as far back as the late 1980s, according to Paul Hoffman, director of the VPN Consortium (www.vpnc.org). Introducing IPsec Vendors initially used proprietary and other forms of encryption with their VPNs. However, to establish a standard way to create interoperable VPNs, many vendors moved to IPsec, which the Internet Engineering Task Force (IETF) adopted in 1998. With IPsec, a computer sends a request for data from a server through a gateway, acting essentially as a router, at the edge of its network. The gateway encrypts the data and sends it over the Internet. The receiving gateway queries the incoming packets, authenticates the sender’s identity and designated network-access level, and if everything checks out, admits and decrypts the information. Both the transmitter and receiver must support IPsec and share a public encryption key for authentication. 17 December 2007 Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F TECHNOLOGY NEWS File and media server Firewall Terminal services Internet SSL encrypted Remote user: traffic Business partner Kiosk user Temporary staff Traveling staff Telecommuter Desktop Decrypted traffic SSL VPN: Authentication Authorization Decryption Integrity check Web proxy Web server E-mail server Figure 1. In an SSL VPN, a remote user logs in to a dedicated Web site to access a company’s network. The user’s browser initiates the session with a corporate server or desktop computer, which downloads the necessary software to the client. The software uses SSL for encrypting the transmitted data. At the corporate site, the VPN system authenticates users, determines what level of network access they should have, and if everything checks out, decrypts the data and sends it to the desired destination. Unlike SSL, IPsec is implemented as a full application installed on the client. And it doesn’t take advantage of existing browser code. IPsec limitations According to Forrester’s Silva, corporate IT departments increasingly need to let remote users connect to enterprise networks, which is challenging with IPsec. The normal practice of configuring IPsec VPNs to allow full access to a network can create vulnerabilities. To avoid this, administrators would have to configure them to permit access only to parts of a network, according to Peter Silva, technical marketing manager for F5 Networks SSL VPNs. IPsec VPNs also have trouble letting certain traffic transverse firewalls, he explained. This isn’t usually a problem, as most companies have the same basic ports open both inbound and outbound. However, it is possible that one company would let traffic out over a port that another doesn’t leave open for inbound data. By contrast, the vast majority of companies have port 80 (dedicated 18 to HTTP traffic) or 443 (dedicated to SSL or HTTPS) open inbound and outbound, so crossing firewalls is rarely a problem for SSL VPNs, which are Web-based. IPsec VPNs are full programs and thus are large, generally 6 to 8 megabytes. This means they download more slowly and don’t always work well on smaller devices. ENTER THE SSL VPN The first SSL VPN vendor was Neoteris, purchased in 2003 by NetScreen, which Juniper bought the next year, according to Juniper’s Campagna. SSL Netscape Communications developed SSL and released the first public version in 1994. The IETF adopted the technology as a standard in 1999, naming it Transport Layer Security. However, most users still call it SSL. The technology, which offers the same encryption strengths as IPsec, has been used largely to secure financial transactions on the Web. In an SSL VPN, a user logs in to a dedicated Web site. The browser initiates the session with the Web server, which downloads the necessary software to the client, generally using either ActiveX or Java controls. Administrators can configure an SSL VPN gateway to conduct additional checks, such as whether the connecting device has the latest security upgrades. During this process, the client and server identify common security parameters, such as ciphers and hash functions, and use the strongest ones they both support. The VPN gateway identifies itself via a digital certificate that includes information such as the name of the trusted authority that issued the certificate, which the client can contact for verification, and the server’s public encryption key. The gateway then sends an encrypted session cookie to the browser to start the communications. To generate the encryption key used for the session, the client encrypts a random number with the server’s public key and sends the result to the server, which decrypts it with a private key. Once the user’s identity is authenticated, an SSL VPN, like an IPsec VPN, allows the level of access granted by company policies for different types of users. Thus, for example, the vice president of human resources would have access to an employee salary database while most other visitors wouldn’t. All major browsers are SSLenabled, so SSL VPNs can work with almost any browser and are thus platform- and operating-systemindependent, said the VPN Consortium’s Hoffman. This makes them more convenient to use, particularly for mobile users, than IPsec VPNs. SSL advantages The mobile or stationary user connects to a company’s SSL VPN by entering a URL in a browser and then presenting login credentials, usually a username and password. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page This begins the process of establishing a secure connection. Basic functionality and implementation. Once the initial connection is made, the Web server downloads controllers that work with the browser’s own code. Thus, the downloads are small, generally between 100 kilobits and 1 megabit. Because they’re so small, they download fast (making them easier for ad hoc use), take up less space on the hard drive, and work better on smaller devices such as cellular phones. Most SSL VPNs use a reverse proxy, which rewrites the content from the Web application and presents it to the browser. Proxy servers present a single interface to users, accelerate the encryption process, perform data compression, and provide an additional layer of security. Gateway boxes are the network element that SSL vendors sell. They streamline both ingoing and outgoing traffic and provide proxying, authentication, and authorization. Authentication and authorization. SSL VPNs check usernames, passwords, and digital certificates to authenticate visitors. The gateways then consult a database to determine the level of network access the user should have. This gateway functionality consolidates the combination of firewalls, extranets, and other technologies previously used to provide authentication. This not only simplifies the process but also reduces the amount of equipment that companies must manage. Challenges Many businesses might not want to change from the IPsec VPNs they’ve spent money on, at least until they have recovered their investments. Users of new SSL VPNs also face a learning curve. Companies need different sets of rules for different users, to provide them with varying degrees of network access. For example, corporate accountants should have access to financial records, but outside visitors should not. If an enterprise doesn’t have such rules in place already, designing and implementing them can require considerable work. In addition, companies must choose among multiple approaches to limiting access, including mapping certain users to parts of a network or building tunnels to specific applications, servers, ports, or filters. SSL VPNs used to be considerably slower than their IPsec-based counterparts. SSL works with TCP, in which data recipients must acknowledge every incoming packet. IPsec, on the other hand, works with the User Datagram Protocol, which is quicker because it doesn’t require acknowledgments. Also, many backbone providers give UDP traffic higher priority than TCP communications. Now, though, SSL VPNs use network optimization and data compression to improve performance. SSL VPNs are also faster than before because they can use the IETF’s relatively new Datagram Transport Layer Security protocol, which runs over UDP, said Cisco product marketing manager Mark Jansen. A BEMaGS F O ver time, SSL VPNs will gain additional capabilities. For example, Forrester’s Silva said they are now able to manage users’ connections and preserve their sessions as, for example, they roam from a Wi-Fi network to a public cellular network and back. IPsec VPNs will still be sufficient for communications between an ITmanaged machine and a network, or for hub-and-spoke communications within a network. Over time, though, in an increasingly mobile world, said Forrester’s Silva, SSL will become the obvious choice for VPNs. ■ Karen Heyman is a freelance technology writer based in Santa Monica, California. Contact her at _________ klhscience@ yahoo.com. Editor: Lee Garber, Computer, l.garber@computer.org ______________ IEEE Computer Society members SAVE 25% on all confer ences sponsor ed by the IEEE Computer Society w w w. c o m p u t e r. o r g / j o i n 19 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F NEWS BRIEFS Robot Adapts to Changing Conditions A US researcher has developed a robot that can adapt to changes to either its environment or its own structure, even if it is damaged. University of Vermont assistant professor Josh Bongard said his Starfish robot doesn’t regenerate a new limb if one is destroyed, like its biological counterpart, but rather adapts its behavior in response to damage and learns to walk in new ways. “The approach is unique in that the robot adapts by creating a computer simulation of itself: how many legs it has, how those legs are arranged, whether the ground over which it’s moving is flat or rocky,” he explained. “By creating such a simulation, it can mentally rehearse ways of walking before executing them.” “This is particularly useful in dan- gerous situations,” he said. “If the robot is perched on the edge of a cliff, it shouldn’t simply try to walk forward and see what happens. It should build an understanding of its immediate environment first.” The technology could be used in disasters or dangerous situations. This work is also important to the creation of self-configuring robots that could change locomotion, for example from crawling to walking, based on various conditions, said Bongard, whose work is funded largely by the US National Science Foundation. Working with Cornell University’s Computational Synthesis Laboratory (CCSL), researchers used 3D printing technology to fabricate the battery-powered robot’s plastic body. Starfish contains a small PC-104 computer with a Pentium 166-MHz The Starfish robot uses sensors and simulation software to adapt to changes to either its environment or its own structure, such as the loss of a limb. Starfish rocks back and forth to enable sensors to determine the nature of the surrounding terrain or the robot’s current physical structure. It then simulates possible movements to determine which will best fulfill its mission. 20 Computer Computer processor, 64 megabits of RAM, and 512 megabits of compact flash memory. It gets a sense of its environment and physical structure by rocking back and forth. This activates jointangle sensors, which determine how the robot is tilting in response to the surrounding terrain or missing parts. The robot collects data on its flash card and uploads the information to an external PC for processing by the simulation software. The external PC sends the robot acquired information about its actions and behavior. Other hardware includes a Diamond DMM-32X-AT dataacquisition board, used for collecting joint-angle sensor data; and a Pontech SV203 servo control board, which drives Starfish’s hinge motors. The researchers program the robot with basic information about its design, such as the mass and shape of its parts. Starfish then builds a virtual model of itself, using Open Dynamics Engine software, which integrates the equations of motion for a connected set of physical objects, such as those that make up the robot. Starfish can thus consider the physical repercussions of a given set of torques applied over time, such as whether a specific motor program— an application that determines how the robot will move—will cause it to go forward or simply shake in place. According to Bongard, his team layered an optimization technique on top of the simulator that lets the robot determine which of its large set of motor programs will best fulfill its mission. “[Other] lines of inquiry that I’m pursuing involve looking at ways to get teams of robots to collectively simulate their surroundings, and then share those simulations,” he added. “In this way, robots could work together to solve tasks that are beyond any one of them.” ■ News Briefs written by Linda Dailey Paulson, a freelance technology writer based in Ventura, California. Contact her at ________________ ldpaulson@yahoo.com. Published by the IEEE Computer Society Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F New Approach Will Speed Up USB I ntel is leading an effort to create a new, much faster version of the Universal Serial Bus standard used for connecting computers to peripherals and other devices. Proponents say faster connections are necessary to transfer the large amounts of dataintensive digital content found today in a reasonable amount of time between computers and devices. To develop SuperSpeed USB technology, Intel has formed the USB 3.0 Promoters Group with members Hewlett-Packard, Microsoft, NEC, NXP Semiconductors, and Texas Instruments. Once they design the standard, the USB Implementers Forum (www. ____ usb.org) will handle tasks such as product-compliance testing and certification, as well as trademark protection and user education, said Jeff Ravencraft, Intel technology strategist and Promoters Group chair. Proponents plan for USB 3.0 to provide 10 times the bandwidth of USB 2.0, which would raise speeds from 480 megabits per second to 4.8 Gbps. The goal is to reduce the wait that consumers currently experience moving rich digital video, audio, and other content between their PCs and portable devices, explained Ravencraft. USB 3.0 users could transfer a 27-gigabyte high-definition movie to a portable media player in 70 seconds, as opposed to the 14 Interface Makes a Computer Desktop Look like a Real Desktop He said the BumpTop software works in the Anand Agrawala has something for computer users background when other applications are open and who like their work in piles rather than files. does not unduly tax the host computer’s CPU. Agrawala, cofounder and CEO of BumpTop, has BumpTop hooks into the operating and file systems created a computer desktop environment that lets so that it accurately reflects when files are created, users place their work in files or piles, represented by moved, or modified. graphics that make the environment look like a real Although the system currently works only with desktop. They can also manipulate and move their Windows, Agrawala said he plans to make it support material and convert it from piles to files and back multiple platforms. again, just as they would with physical items. Agrawala said he is still working on a business model This enables them to function more comfortably by for BumpTop and is developing an alpha version of the letting them handle their computer documents in the system. He plans to release a finished product in the same way they work with their physical documents, near future. ■ explained Agrawala, who created the new interface during his master’s degree studies at the University of Toronto. The interface also gives the desktop a third dimension, which provides more information density than conventional environments. To add functionality to the standard desktop, Agrawala used gaming technology. For example, the system employs the Ageia PhysX physics engine to handle the graphics in a stable manner. BumpTop uses a gaming engine to run the complex mathematics that controls the movement of documents on the desktop and renders them properly. “Say you want to move an object by throwing it into the corner. The math determines Anand Agrawala (right), cofounder and CEO of BumpTop, demonstrates his new computer how fast an object will slow down desktop environment, which lets users place their work in files or piles, shown by graphics and the friction that will come into that make the interface look like a real desktop. Users can also manipulate and move their play,” Agrawala explained. material, just as they would with physical items. 21 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F NEWS BRIEFS minutes required with USB 2.0. The new technology might even allow a device to send high-definition signals to a TV, noted analyst Carl D. Howe, a principal with consultancy Blackfriars Communications. USB 3.0 would also work with USB drives, camcorders, external hard drives, MP3 players, portable telephones, and other flash-capable portable devices, Ravencraft said. It will be backward compatible with earlier USB versions, using the same connectors and programming models, and will support multiple communications technologies, he added. USB 3.0 will improve performance in part by using two channels to sep- arate data transmissions and acknowledgements, thereby letting devices send and receive simultaneously, rather than serially. Unlike USB 2.0, the new technology won’t continuously poll devices to find out if they have new data to send, which should reduce power consumption. USB 3.0 will also add quality-ofservice capabilities to guarantee specified levels of resources to highpriority traffic. Ravencraft declined to provide more detail, explaining that “the specification is still being defined.” Intel expects the final standard to be finished by mid-2008 and to appear in peripherals and devices in 2009 or 2010. USB’s main competitors among external bus technologies are FireWire (IEEE 1394), which has a new standard that runs at 3.2 Gbps; and eSATA (external serial advanced technology attachment), which offers up to 3 Gbps now and a planned 6 Gbps by mid-2008. Said Howe, “My guess is that video recorders and the like may go with FireWire, which can run … without imposing the big computational load that USB does. And there are other systems, like wireless USB, that promise similar speeds but without the wires.” ■ Standardization Comes to Virtualization S everal companies are cooperating to develop virtualization standards. This would enable products from virtualization firms to work together, an important development for both vendors and users of the increasingly popular technology. Virtualization software enables a single computer to run multiple operating systems simultaneously, via the use of virtual machines (VMs). The products let companies use a single server for multiple tasks that would normally have to run on multiple servers, each working with a different OS. This reduces the number of servers a company needs, thereby saving money and using resources efficiently. Currently, different vendors’ VMs use their own formats and won’t necessarily run on other virtualization vendors’ platforms, explained Simon Crosby, chief technology officer with virtualization vendor XenSource. Different vendors’ products also use separate formats for storing VMs on a user’s hard drive. To overcome this, virtualization vendors Microsoft, VMware, and XenSource, along with server makers Dell, Hewlett-Packard, and IBM and the Distributed Management 22 Task Force are working on the Open Virtual Machine Format. The DMTF is an international industry organization that develops management standards. OVF defines an XML wrapper that encapsulates multiple virtual machines and provides a common interface so that the VMs can run on any virtualization system that supports OVF. A vendor could deliver an OVF-formatted VM to customers who could then deploy it on the virtualization platform of their choice. For example, a VM packaged for a VMware system could run on a XenSource system. OVF utilizes existing tools to combine multiple virtual machines with the XML wrapper, which gives the user’s virtualization platform a package containing all required installation and configuration parameters for the VMs. This lets the platforms run any OVF-enabled VM. Vendors are cooperating in this effort because users demand interoperability among different products for running multiple VMs on one computer, noted Laura DiDio, vice president and research fellow for Yankee Group, a market research firm. More- over, interoperability would increase the overall use of virtualization. DiDio said Yankee Group research shows 96 percent of all organizations plan to adopt virtualization, and one-third of them will use multiple virtualization products to get the exact array of capabilities they need. Standards-based interoperability will thus be critical, she explained. In addition to enabling interoperability, OVF provides security by attaching digital signatures to virtual machines. This verifies that the OVF file actually came from the indicated source and lets users’ systems determine whether anyone has tampered with the VMs. There is no firm timeline for adopting and implementing the OVF yet. “I’m optimistic that we’ll see something in the first half of 2008,” said Winston Bumpus, DMTF president and VMware’s director of standards architecture. The task force is also working on standards for other aspects of virtualization, added Bumpus. ■ Editor: Lee Garber, Computer, l.garber@computer.org ______________ Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page PURPOSE: The IEEE Computer Society is the world’s largest association of computing professionals and is the leading provider of technical information in the field. MEMBERSHIP: Members receive the monthly magazine Computer, discounts, and opportunities to serve (all activities are led by volunteer members). Membership is open to all IEEE members, affiliate society members, and others interested in the computer field. COMPUTER SOCIETY WEB SITE: www.computer.org OMBUDSMAN: To check membership status or report a change of address, call the IEEE Member Services toll-free number, +1 800 678 4333 (US) or +1 732 981 0060 (international). Direct all other Computer Society-related questions—magazine delivery or unresolved complaints—to ___________ help@computer.org. CHAPTERS: Regular and student chapters worldwide provide the opportunity to interact with colleagues, hear technical experts, and serve the local professional community. AVAILABLE INFORMATION: To obtain more information on any of the following, contact Customer Service at +1 714 821 8380 or +1 800 272 6657: • • • • • • • • • Membership applications Publications catalog Draft standards and order forms Technical committee list Technical committee application Chapter start-up procedures Student scholarship information Volunteer leaders/staff directory IEEE senior member grade application (requires 10 years practice and significant performance in five of those 10) PUBLICATIONS AND ACTIVITIES Computer. The flagship publication of the IEEE Computer Society, Computer, publishes peer-reviewed technical content that covers all aspects of computer science, computer engineering, technology, and applications. Periodicals. The society publishes 14 magazines, 9 transactions, and one letters. Refer to membership application or request information as noted above. Conference Proceedings & Books. Conference Publishing Services publishes more than 175 titles every year. CS Press publishes books in partnership with John Wiley & Sons. Standards Working Groups. More than 150 groups produce IEEE standards used throughout the world. Technical Committees. TCs provide professional interaction in over 45 technical areas and directly influence computer engineering conferences and publications. Conferences/Education. The society holds about 200 conferences each year and sponsors many educational activities, including computing science accreditation and certification. A BEMaGS F EXECUTIVE COMMITTEE President: Michael R. Williams* President-Elect: Rangachar Kasturi* Past President: Deborah M. Cooper* VP, Conferences and Tutorials: Susan K. (Kathy) Land (1ST VP)* VP, Electronic Products and Services: Sorel Reisman (2ND VP)* VP, Chapters Activities: Antonio Doria* VP, Educational Activities: Stephen B. Seidman† VP, Publications: Jon G. Rokne† VP, Standards Activities: John Walz† VP, Technical Activities: Stephanie M. White* Secretary: Christina M. Schober* Treasurer: Michel Israel† 2006–2007 IEEE Division V Director: Oscar N. Garcia† 2007–2008 IEEE Division VIII Director: Thomas W. Williams† 2007 IEEE Division V Director-Elect: Deborah M. Cooper* Computer Editor in Chief: Carl K. Chang† Executive Director: Angela R. Burgess† * voting member of the Board of Governors † nonvoting member of the Board of Governors BOARD OF GOVERNORS Term Expiring 2007: Jean M. Bacon, George V. Cybenko, Antonio Doria, Richard A. Kemmerer, Itaru Mimura, Brian M. O’Connell, Christina M. Schober Term Expiring 2008: Richard H. Eckhouse, James D. Isaak, James W. Moore, Gary McGraw, Robert H. Sloan, Makoto Takizawa, Stephanie M. White Term Expiring 2009: Van L. Eden, Robert Dupuis, Frank E. Ferrante, Roger U. Fujii, Ann Q. Gates, Juan E. Gilbert, Don F. Shafer EXECUTIVE STAFF Executive Director: Angela R. Burgess Associate Executive Director: Anne Marie Kelly Associate Publisher: Dick Price Director, Administration: Violet S. Doan Director, Finance and Accounting: John Miller COMPUTER SOCIETY OFFICES Washington Office. 1730 Massachusetts Ave. NW, Washington, DC 20036-1992 Phone: +1 202 371 0101 • Fax: +1 202 728 9614 Email: ___________ hq.ofc@computer.org Los Alamitos Office. 10662 Los Vaqueros Circle, Los Alamitos, CA 90720-1314 Phone: +1 714 821 8380 Email: __________ help@computer.org Membership and Publication Orders: Phone: +1 800 272 6657 • Fax: +1 714 821 4641 Email: ___________ help@computer.org Asia/Pacific Office. Watanabe Building, 1-4-2 Minami-Aoyama, Minato-ku, Tokyo 107-0062, Japan Phone: +81 3 3408 3118 • Fax: +81 3 3408 3553 Email: ____________ tokyo.ofc@computer.org IEEE OFFICERS President: Leah H. Jamieson President-Elect: Lewis Terman Past President: Michael R. Lightner Executive Director & COO: Jeffry W. Raynes Secretary: Celia Desmond Treasurer: David Green VP, Educational Activities: Moshe Kam VP, Publication Services and Products: John Baillieul VP, Regional Activities: Pedro Ray President, Standards Association: George W. Arnold VP, Technical Activities: Peter Staecker IEEE Division V Director: Oscar N. Garcia IEEE Division VIII Director: Thomas W. Williams President, IEEE-USA: John W. Meredith, P.E. Next Board Meeting: 16 May 2008, Las Vegas, NV, USA revised 11 Oct. 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F COMPUTING PRACTICES Examining the Challenges of Scientific Workflows Workflows have emerged as a paradigm for representing and managing complex distributed computations and are used to accelerate the pace of scientific progress. A recent National Science Foundation workshop brought together domain, computer, and social scientists to discuss requirements of future scientific applications and the challenges they present to current workflow technologies. Yolanda Gil and Ewa Deelman University of Southern California Mark Ellisman University of California, San Diego Thomas Fahringer University of Innsbruck Geoffrey Fox and Dennis Gannon Indiana University Carole Goble Manchester University Miron Livny University of WisconsinMadison Luc Moreau University of Southampton Jim Myers National Center for Supercomputing Applications 24 Computer Computer S ignificant scientific advances are increasingly achieved through complex sets of computations and data analyses. These computations may comprise thousands of steps, where each step might integrate diverse models and data sources that different groups develop. The applications and data might be also distributed in the execution environment. The assembly and management of such complex distributed computations present many challenges, and increasingly ambitious scientific inquiry is continuously pushing the limits of current technology. Workflows have recently emerged as a paradigm for representing and managing complex distributed scientific computations, accelerating the pace of scientific progress.1-6 Scientific workflows orchestrate the dataflow across the individual data transformations and analysis steps, as well as the mechanisms to execute them in a distributed environment. Each step in a workflow specifies a process or computation to be executed (for instance, a software program or Web service). The workflow links the steps according to the data flow and dependencies among them. The representation of these computational workflows contains many details required to carry out each analysis step, including the use of specific execution and storage resources in distributed environments. Figure 1 shows an example of a high-level workflow developed within the context of an earthquake science application, CyberShake (www.scec.org), which generates shake maps of Southern California.7 Workflow systems exploit these explicit representations of complex computational processes at various levels of abstraction to manage their life cycle and automate their execution. In addition to automation, workflows can provide the information necessary for scientific reproducibility, result derivation, and result sharing among collaborators. By providing automation and enabling reproducibility, they can accelerate and transform the scientific-analysis process. Workflow systems have demonstrated these capabilities in a variety of applications where workflows comprising thousands of components processed large, Published by the IEEE Computer Society 0018-9162/07/$25.00 © 2007 IEEE Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F XYZInputFile_1 SortedSRLFile_1 File_SeisParamVals1 FileCollection_RVM1 CollOfCollection_RuptureVariations1 Node_GenMetaforPeakValCal PreSGTInputFile_1 Node_PreSGT Node_PreSGT_Partial PreSGTFileCollectionDescription_1 Node_BoxNameCheck FileCollection_SGTFileDescriptions1 RupturesForSite_1 IN3DFile_1 Node_pvm1_chk1 FileCollection_SeisMetadataforPVC1 SGTFilesForSiteDir_1 Node_CreateCVMFilesDesc GRDFile_1 EdgeFileCollection_1 CMVFileCollectionDescription_1 PreSGTFileForSite_1 CollOfCollection_SGTCollection1 Node_SynthSGT_Collection Node_XYZGRD CVMFileForSite_1 Node_FD_GRID_CVM IN3DFile_2 Node_pvm1_chk2 SGTFilesForSiteDir_2 FileCollection_Seismogram1 Node_PeakValCal_Collection FileCollection_SAOutputFile1 Figure 1. A visual representation of a high-level workflow developed within the context of an earthquake science application. Double-lined nodes indicate computations that the system will parallelize automatically. distributed data sets on high-end computing resources. Some workflow systems are deployed for routine use in scientific collaboratories—virtual entities that allow scientists to collaborate with each other across organizations and physical locations. Figure 2 shows an image of the Orion Nebula that the Montage8 application produced. Montage uses workflow technologies9 to generate science-grade mosaics of the sky. Researchers recently used such mosaics to verify a bar in the M31 galaxy.10 Much research is under way to address issues of creation, reuse, provenance tracking, performance optimization, and reliability. However, to fully realize the promise of workflow technologies, we must meet many additional requirements and challenges. Scientific applications are driving workflow systems to examine issues such as supporting dynamic event-driven analyses, handling streaming data, accommodating interaction with users, intelligent assistance and collaborative support for workflow design, and enabling result sharing across collaborations. As a result, we need a more comprehensive treatment of workflows to meet the long-term requirements of scientific applications. The National Science Foundation’s 2006 Workshop on Challenges of Scientific Workflows brought together domain, computer, and social scientists to discuss requirements of future scientific applications and the challenges they present to current workflow technologies. As part of the workshop, we examined application requirements, workflow representations, dynamic workflows, and system-related challenges. Figure 2.The Montage application uses workflow technologies to generate science-grade mosaics of the sky. 25 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F APPLICATION REQUIREMENTS tion with the analysis results, reproducing important discoveries involving complex computations can be impractical or even impossible. To support reproducibility, workflow-management systems must capture and generate provenance information as a critical part of the workflow-generated data. Collaborations Workflow-management systems must also consume the Combining distributed data, computation, models, provenance information associated with input data and and instruments at unprecedented scales can enable associate that information with the resulting data prodtransformative research. The analysis of large amounts ucts. Systems must associate and store provenance with of widely distributed data is becoming commonplace. the new data products and contain enough details to This data, and the experimental apparatus or simula- enable reproducibility. tion systems that produce it, typically Scientists also need interoperable, belong to collaborations rather than persistent repositories of data and individuals. Within these collaboraanalysis definitions, with linkage to Many disciplines benefit tions, various individuals are responopen data and publications, as well from the use of sible for different aspects of data as to the algorithms and applicaacquisition, processing, and analysis, tions used to transform the data. workflow-management and entire projects often generate Workflow systems must complesystems to automate publications. Such environments ment existing data repositories with computational activities. demand tools that can orchestrate provenance and metadata repositothe steps of scientific discovery and ries that enable the discovery of the bridge the differing expertise of colworkflows and application compolaboration members. nents used to create the data. Two important concerns Many disciplines benefit from the use of workflow- for scientists in these highly collaborative endeavors management systems to automate such computational are credit assignment and recognition of individual activities, including astronomy, biology, chemistry, envi- contributions. ronmental science, engineering, geosciences, medicine, physics, and social sciences. Flexible environments The scientific community perceives that workflows are Systems must be flexible in terms of supporting both important in accelerating the pace of scientific discov- common analyses that many scientists perform, as well eries. Today, complex scientific analyses increasingly as unique analyses. Researchers should find it easy to require tremendous amounts of human effort and man- set up and execute routine analyses based on common ual coordination. Thus, researchers need more effective cases. At the same time, individual scientists should be tools to prevent being inundated by the ever-growing able to steer the system to conduct unique analyses and data and associated computational processing tasks. create novel workflows with previously unseen combinations and configurations of models. Reproducibility From an operational perspective, there’s a need to proThe NSF workshop participants identified repro- vide secure, reliable, and scalable solutions. Scientists ducibility of scientific analyses and processes as an must trust that their input and output data is secure and important application requirement. Reproducibility is free from inappropriate data access or malicious manipat the core of the scientific method, enabling scientists to ulation. Current infrastructure must incorporate trust evaluate the validity of each other’s hypotheses and pro- and reputation systems for data providers. viding the basis for establishing known truths. Finally, scientists need easy-to-use tools that provide Reproducibility requires rich provenance information intelligent assistance for such complex workflow capaso researchers can repeat techniques and analysis meth- bilities. Automation of low-level operational aspects of ods to obtain scientifically similar results. workflows is a key requirement. Success will depend on Today, reproducibility for complex scientific applica- interaction modalities that hide unnecessary complexitions is virtually impossible. Many scientists are ties and speak the scientist’s language. involved, and the provenance records are highly fragmented, existing in e-mails, wiki entries, database SHARED WORKFLOW DESCRIPTIONS queries, journal references, codes, and other sources for Given the broad practice and benefits of sharing communication. All this information, often stored in instruments, data, computing, networking, and many various locations and forms, must be appropriately other science products and resources, why don’t indexed and made available for referencing. Without researchers widely capture and share scientific computracking and integrating these crucial bits of informa- tations and processes as well? Given the exponential growth in computing, sensors, data storage, network, and other performance elements, why is the growth of scientific data analysis and understanding not proportional? 26 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Process sharing rich semantic representations of requirements and conScientists have always relied on technology to share straints on workflow models and components. With information about experiments, from pen and paper to semantic descriptions of the data format and type requiredigital cameras, e-mail, the Web, and computer soft- ments of a component, it’s possible to incorporate autoware. Workflow description and execution capabilities mated reasoning and planning capabilities that could offer a new way of sharing and managing information automatically add data conversion and transformation to electronically capture full processes and share them steps. Similarly, rich descriptions of the execution requirefor future reference and reuse. ments of each workflow component would enable autoThis new way of sharing information—agreeing on mated resource selection and dynamic optimizations. processes’ semantics and the infrastructure to support Levels of description. Abstractions would let scientheir execution—continues the historic push for making tists identify what levels of description are useful to share representations explicit and actionable and reducing the in their workflows, and they could package such descripbarriers to coordination. We should tions as a self-contained sharable encourage scientists to bring workobject that other scientists could then Workflow representations flow representations to their practices refine and instantiate. We need refineand share the descriptions of their sciment and abstraction capabilities for must incorporate rich entific analyses and computations in all first-class entities that workflow information about analysis ways that are as formal and explicit systems must manipulate: workflow processes to support as possible. However, no commonly scripts (regarded as specifications of accepted and sufficiently rich reprefuture execution), provenance logs discovery, creation, sentations exist in the scientific com(descriptions of process and data hismerging, and execution. tory), data, and metadata. There’s munity. relevant work in related fields of computer science, such as refinement Representations Workflow representations must accommodate scien- calculi, model-driven architectures, and semantic modtific process descriptions at multiple levels. For instance, eling, but researchers haven’t applied these techniques domain scientists might want a sophisticated graphical widely to scientific workflows, which are potentially interface for composing relatively high-level scientific or large scale, might involve multiple technologies, and mathematical steps, whereas the use of a workflow lan- must operate on heterogeneous systems. guage and detailed specifications of data movement and The sophistication of required descriptions depends job execution steps might concern computer scientists. on the workflow capabilities needed. For example, a To link these views and provide needed capabilities, workflow that adapts dynamically to changes in enviworkflow representations must include rich descriptions ronment or data values requires formal and comprethat span abstraction levels and include models of how hensive descriptions to enable automatic adaptation. to map between them. Further, to support the end-to- Even for a human to make choices related to making end description of multidisciplinary, community-scale changes to a workflow would require access to a broad research, we need definitions of workflow and prove- variety of descriptions. nance that are broad enough to describe workflows-ofworkflows that are linked through reference data, the Scientific versus business workflows scientific literature, and manual processes in general. Understanding the differences between scientific Other important and necessary dimensions of abstrac- workflows and practices and those used in business tion are experiment-critical versus non-experiment-crit- could yield useful insights. On the one hand, scientific ical representations, where the former refers to scientific and business workflows aren’t obviously distinguishissues and the latter is more concerned with operational able, since both might share common important charmatters. acteristics. Indeed, the literature contains examples of workflows in both domains that are data-intensive and Abstractions highly parallel. On the other hand, scientific research Workflow representations must incorporate rich infor- requires flexible design and exploration capabilities that mation about analysis processes to support discovery, appear to depart significantly from the more prescriptive creation, merging, and execution. These activities will use of workflows in business. Workflows in science are become a natural way to conduct experiments and share a means to support detailed scientific discourse as well scientific methodology within and across scientific as a way to ensure repeatable processes. communities. Another distinctive issue of scientific workflows is the Automation. Wherever possible, workflow represen- variety and heterogeneity of data within a single worktations need to support automation of the workflow cre- flow. For example, a scientific workflow might involve ation and management processes. This capability requires numeric and experimental data in proprietary formats 27 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page (such as those used for raw data that scientific instruments involved in a process produce), followed by processed data resulting in descriptions related to scientific elements, leading to textual, semistructured, and structured data, and formats used for visual representation. To clarify the research issues in developing scientific workflow capabilities, the community needs to identify where there are real differences between scientific and business activities, beyond domain-specific matters. It’s important to balance the desire for sharing workflow information against the dangers of premature standardization efforts that might constrain future requirements and capabilities. A BEMaGS F challenge is to develop mechanisms to create, manage, and capture dynamic workflows to allow for reproducibility of significant results. Scientific practice will routinely give rise to dynamic workflows that base decisions about subsequent steps on the latest available information. Researchers might need to dynamically design a workflow to look at the initial steps’ results before making a decision on carrying out later analysis steps. For example, examining the results of an image’s initial preprocessing might require subsequent steps to look at specific areas that preprocessing identified. External events A dynamic workflow could also result from an external event changing the workflow’s Most scientific activity consists of basic structure or semantics. For The management exploration of variants and experiexample, in severe-storm prediction, mentation with alternative settings, data-analysis computations might of dynamic which would involve modifying search for patterns in radar data. workflows is workflows to understand their Depending upon the specific pattern complex due to effects and provide a means for of events, enacting different branches explaining those effects. Hence, an of a storm-prediction workflow their evolution important challenge in science is repmight require significant computaand life cycles. resentation of workflow variants, tional resources on-demand. which aims at understanding the In this case, the workflow must impact that a change has on the adapt to changes in storm intensity resulting data products as an aid to scientific discourse. or resource availability. Some experimental regimens While acknowledging that sharing representations is might draw on workflows that are heuristic or employ important to the scientific process, the workshop group untried activities, thus these workflows might break recognized that workflows must accommodate multi- down or fail during their execution, necessitating fault ple collaboration and sharing practices. In some cases, diagnosis and repair. Two workflows could also affect it’s suitable to share workflows, but not data. In other each other by sharing results, being classified as dynamic cases, scientists want to share an abstract description as they respond to events arising in each other’s execuof the scientific protocol without actually communi- tion. Finally, some scientific endeavors are large scale. They cating details, parameters, and configurations, which are their private expertise. In other situations, a involve large teams of scientists and technicians, and description of a specific previous execution (prove- they engage in experimental methods or procedures that nance) is desirable, with or without providing execu- take a long time to complete and require human intertion details. vention and dynamic steering throughout the process. For example, an astrophysical study of deep-space phenomena might require the use and coordination of mulDYNAMIC WORKFLOWS How can workflows support both the exploratory tiple observation devices operating in different spaces, nature of science and the dynamic processes involved in capturing data at different frequencies or modalities. scientific analysis? Workflow variants Workflow life cycle Changing context and infrastructure Given that both the user’s experimental context and the distributed infrastructure that the workflows operate over are in flux, the notion of static workflows is an odd one. The vision of supporting dynamic, adaptive, and user-steered workflows is to enable and accelerate distributed and collaborative scientific methodology via rapid reuse and exploration accompanied by continuous adaptation and improvement. Reproducibility becomes ever more elusive in this kind of setting. The 28 The management of dynamic workflows is complex due to their evolution and life cycles. As Figure 3 shows, there’s no beginning or end to the life-cycle process of a workflow—scientists can start at any point and flow through the figure in any direction. They might build or assemble a workflow, refine one that a shared repository has previously published, run their design, evolve it, run it again, share fragments of it as they go along, find other fragments they need, run it a few more times, and learn from the protocol they’re developing. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page They might settle on the workflow and run it many times, learning from the results produced, or they might run it just once because that’s all they need. While running, the workflows could adapt to external events and user steering. The results of the whole activity feed into the next phases of investigation. The user is ultimately at the center, interacting with the workflows and interpreting the outcomes. Supporting scientists in complex exploratory processes involving dynamic workflows is an important challenge. Researchers will need to design a human-centered decision-support system that accommodates the information needs of a scientist tracking and understanding such complex processes. The workflow will need appropriate user interfaces that enable scientists to browse/traverse, query, recapitulate, and understand this information. Simplifying the exploratory process also requires novel and scalable means for scientists to manipulate the workflows, explore slices of the parameter space, and compare the results of different configurations. Learning workflow patterns An interesting direction for future research explores the question of how to improve, redesign, or optimize workflows through data mining of workflow life-cycle histories to learn successful (and unsuccessful) workflow patterns and designs, and assist users in following (or avoiding) them. Researchers can extract one kind of pattern from successful execution trails and use the information to build recommendation systems. For example, if a model M is added, the system could suggest additional models that other people often use together with M in a workflow, or suggest values commonly used for the parameters in the model. Researchers could extract another kind of pattern from unsuccessful trails. These patterns can, for example, help identify incompatible parameter settings, unreliable servers or services, or gross inefficiencies in resource usage. Researchers can subsequently analyze, reenact (reproduce), and validate workflow patterns in order to facilitate their reuse, continuous improvement, and redeployment into new locations or settings. SYSTEM-LEVEL WORKFLOW MANAGEMENT Given the continuous evolution of infrastructure and associated technology, how can we ensure reproducibility of computational analyses over a long period of time? Engineering reproducibility A key challenge in scientific workflows is ensuring engineering reproducibility to enable the reexecution of analyses, and the replication of results. Scientific reproducibility implies that someone can follow the general methodology, relying on the same initial data, and obtain equivalent results. Engineering reproducibility requires more knowledge of the data manipulations, of A BEMaGS F Build and refine Learn Shared repositories Share Run Figure 3. A view of the workflow life cycle, where the processes of workflow generation, sharing, running, and learning are continuous. the actual software and execution environment (hardware, specific libraries), to replicate the results bit-bybit. Researchers need the former capability when they want to validate each other’s hypotheses, whereas the latter is beneficial when they find unusual results or errors and need to trace and understand them. System stability Providing a stable view in spite of continuous technology and platform changes at the system level will be challenging. Researchers must design the underlying execution system to provide a stable environment for the software layers managing the high-level scientific process. It must be possible to reexecute workflows many years later and obtain the same results. This requirement poses challenges in terms of creating a stable layer of abstraction over a rapidly evolving infrastructure while providing the flexibility needed to address evolving requirements and applications and to support new capabilities. To provide consistent and efficient access to resources, resource management must consider both physical resources (computers, networks, and data servers) and logical resources (data repositories, programs, application components, and workflows). Uniform interfaces should inform both. Enhancing resource descriptions with semantic annotations can enable easier, more organized, and possibly automated provisioning, provenance, configuration, and deployment of new resources. Extending current information services with meaningful semantic descriptions of resources should allow for semiautomatic discovery, brokering, and negotiation. 29 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Dynamic configuration and life-cycle management of resources should minimize human interaction. Researchers have made some efforts to provide semiautomatic discovery and brokering of physical resources and management of software components that might become part of scientific-workflow environments. However, there’s still much opportunity for improvements, since most existing systems require manual or semimanual deployment of software components and force application builders to hard-code software component locations on specific resources into their workflows. becoming large as the quantities of data operated on become larger. As workflows scale from 1,000 to 10,000 and perhaps 1 million or more tasks, researchers might need new techniques to represent sets of tasks, manage those tasks, dispatch tasks efficiently to resources, monitor task execution, detect and deal with failures, and so on. A second important scaling dimension is the number of workflows. Particularly in large communities, many users might submit many workflows at once. If these workflows compete for resources or otherwise interact, the runtime environment needs appropriate supporting Quality of service mechanisms to arbitrate among competing demands. Workflow end users frequently need to specify qualityA third scaling dimension concerns the number of of-service (QoS) requirements. The resources involved. Ultimately, we underlying runtime environment can imagine workflows running on Issues of scale will should then guarantee, or at least millions of data and computing maintain, these requirements on a resources (indeed, some systems such increasingly require best-effort basis. However, current as SETI@home already operate at advances over the systems are mostly restricted to bestthat scale). A fourth scaling dimenstate of the art, effort optimizations for time-based sion concerns the number of particcriteria such as reducing overall exeipants. In a simple case, a single user and they occur in cution time or maximizing bandprepares and submits a workflow. In multiple dimensions. width. Researchers must address a more complex case, many particiseveral problems to overcome current pants might help define the worklimitations. flow, contributing relevant data, We must extend QoS parameters beyond time-based managing its execution, and interpreting results. criteria to cover other important aspects of workflow We need to provide new infrastructure services to supbehavior such as responsiveness, fault tolerance, secu- port workflow management. Some of these services are rity, and costs. To provide a basis for interoperable analogous to existing data management and informaworkflow environments or services, this effort will tion services, such as workflow repositories and regrequire collaborative work on the definition of QoS istries. Other novel services will be concerned with parameters that scientists can widely accept. workflows as active processes and the management of Coping with multicriteria optimization or planning their execution state. might require radically changing current optimization and planning approaches. Many systems exist for sin- Infrastructure constraints gle or bi-criteria optimization, but few systems tackle There’s a perceived tension between workflow multicriteria optimization problems. There’s no ready- research challenges and the constraints that existing proto-use methodology that can deal with this problem in duction-quality infrastructures impose. Shared infra____ an efficient and effective way; thus, many opportunities structures such as the Open Science Grid (www. for research exist. opensciencegrid.org), ______________ the TeraGrid (www.teragrid.org), Reservation mechanisms will be an important tool in and NMI (www.nsf-middleware.org) provide widely developing runtime environment support for QoS. Both used and well-tested capabilities to build on. These sysimmediate and advance reservations can make the tem-level infrastructure layers are designed to be prodynamic behavior of infrastructures more predictable, duction quality, but out of necessity haven’t been an important prerequisite to guarantee QoS parameters designed to address workflows’ specific requirements. such as responsiveness and dependability. Moreover, Rather, they aim to meet a broader research commuadvance reservation can also simplify the scheduling of nity’s needs. It’s unlikely that we can make commitments by selecting particular architectures or implementations workflow tasks to resources. at the workflow layers of shared cyberinfrastructure. Scaling We must explore alternative architectures to understand Challenging issues of scale arise in workflow execu- design tradeoffs in different contexts. Examples include tion. These issues will increasingly require advances over the state of the art, and they occur in multiple • workflows designed and tested on a desktop and run dimensions. with larger data in a cluster, First, in many disciplines, individual workflows are • workflows to handle streaming data, 30 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page • event-driven workflow-management engines, and • architectures centered on interactivity. At the same time, we could design these architectures to be interoperable and compatible, where feasible, with some overall end-to-end, multilevel framework. Followon discussions and workshops to understand and address these issues will be extremely beneficial. RECOMMENDATIONS Workflows provide a formal specification of the scientific-analysis process from data collection, through analysis, to data publication. We can view workflows as recipes for cyberinfrastructure computations, providing a representation describing the end-to-end processes involved in carrying out heterogeneous interdependent distributed computations. Once scientists capture this process in declarative workflow structures, they can use workflow-management tools to accelerate the rate of scientific progress by creating, merging, executing, and reusing these processes. By assisting scientists in reusing well-known and common practices for analyses, complex computations will become a daily commodity for use in scientific discovery. As scientists conduct experiments in neighboring disciplines, cross-disciplinary scientific analyses will become commonplace. The NSF workshop participants made the following recommendations: • Support basic research in computer science to create a science of workflows. • Make explicit workflow representations that capture scientific analysis processes at all levels the norm when performing complex distributed scientific computations. • Integrate workflow representations with other forms of scientific record. • Support and encourage cross-disciplinary projects involving relevant areas of computer science as well as domain sciences with distinct requirements and challenges. • Provide long-term, stable collaborations and programs. • Define a road map to advance the research agenda of scientific workflows while building on existing cyberinfrastructure. • Coordinate between existing and new projects on workflow systems and interoperation frameworks for workflow tools. • Hold follow-up, cross-cutting workshops and meetings and encourage discussions between subdisciplines of computer science. Scientists view workflows as key enablers for reproducibility of experiments involving large-scale compu- A BEMaGS F tations. Reproducibility is ingrained in the scientific method, and there’s concern that without this ability, scientists will reject cyberinfrastructure as a legitimate means for conducting experiments. Representing scientific processes with enough fidelity and flexibility will be a key challenge for the research community. Recognizing that science has an exploratory and evolutionary nature, workflows need to support dynamic and interactive behavior. Thus, workflow systems need to become more dynamic and amenable to steering by users and be more responsive to changes in the environment. W orkflows should become first-class entities in the cyberinfrastructure architecture. For domain scientists, they’re important because workflows document and manage the increasingly complex processes involved in exploration and discovery through computation. For computer scientists, workflows provide a formal and declarative representation of complex distributed computations that must be managed efficiently through their life cycle from assembly, to execution, to sharing. ■ Acknowledgments The NSF sponsored this workshop under grant #IIS0629361. We thank Maria Zemankova, program manager of the Information and Intelligent Systems Division, for supporting the workshop and contributing to the discussions. We also thank the workshop attendees (http://vtcpc.isi.edu/wiki/index.php/Participants) for their contributions. References 1. E. Deelman and Y. Gil, eds. Final Report of NSF Workshop on Challenges of Scientific Workflows, Nat’l Science Foundation; http://vtcpc.isi.edu/wiki/images/b/b2/NSFWorkshopFlyer-final. pdf. __ 2. E.Deelman and I. Taylor, eds., J. Grid Computing, special issue on scientific workflows, vol. 3, no. 3-4, Sept. 2005. 3. E. Deelman, Z. Zhao, and A. Belloum, eds., Scientific Programming J., special issue on workflows to support large-scale science, vol. 14, no. 3-4, 2006. 4. G. Fox and D. Gannon, eds., Concurrency and Computation: Practice and Experience, special issue on workflow in grid systems, vol. 18, no. 10, Aug. 2006. 5. B. Ludaescher and C. Goble, eds., SIGMOD Record, special issue on scientific workflows, vol. 34, no. 3, Sept. 2005; www. ___ sigmod.org/record/issues/0509/index.html. 6. I.J. Taylor et al., eds., Workflows for e-Science: Scientific Workflows for Grids, Springer-Verlag, 2006. 7. Y. Gil et al., “Wings for Pegasus: Creating Large-Scale Scientific Applications Using Semantic Representations of Computational Workflows,” Proc. Conf. Innovative Applications 31 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page of Artificial Intelligence (IAAI), 2007, pp. 1767-1774; ____ http:// dblp.uni-trier.de/rec/bibtex/conf/aaai/GilRDMK07. _______________________________ 8. G.B. Berriman et al., “Montage: A Grid-Enabled Engine for Delivering Custom Science-Grade Mosaics On Demand,” Proc. SPIE, vol. 5493, SPIE, 2004, pp. 221-234. 9. E. Deelman et al., “Pegasus: A Framework for Mapping Complex Scientific Workflows onto Distributed Systems,” Scientific Programming J., vol. 13, 2005, pp. 219-237; ____ http:// vtcpc.isi.edu/wiki/images/f/f4/Pegasus.doc. __________________________ 10. R.L. Beaton et al., “Unveiling the Boxy Bulge and Bar of the Andromeda Spiral Galaxy,” Astrophysical J. Letters (in submission). Yolanda Gil is the associate division director for research of the Intelligent Systems Division at the University of Southern California’s Information Sciences Institute (ISI) and a research associate professor in the Computer Science Department. She cochaired the NSF’s Workshop on Challenges of Scientific Workflows. Her research interests include intelligent interfaces for knowledge-rich problem solving. Gil received a PhD in computer science from Carnegie Mellon University. She is a member of the Association for the Advancement of Artificial Intelligence. Contact her at gil@isi.edu. _______ Ewa Deelman is a project leader in the Advanced Systems Division at USC’s ISI and cochaired the NSF’s Workshop on Challenges of Scientific Workflows. Her main research interest is scientific workflow management in distributed environments. Deelman received a PhD in computer science from Rensselaer Polytechnic Institute. She is a member of the IEEE Computer Society. Contact her at _______ deelman@ isi.edu. _____ Mark Ellisman is the director of the Center for Research in Biological Systems and National Center for Microscopy and Imaging Research and a professor of neurosciences and bioengineering at the University of California, San Diego. His research interests include the molecular and cellular basis of nervous system function as well as the use of advanced imaging and information technologies in brain research. Ellisman received a PhD in molecular, cellular, and developmental biology from the University of Colorado, Boulder. He is a founding fellow of the American Institute of Medical and Biological Engineering. Contact him at ________ mellisman@ ucsd.edu. ______ Thomas Fahringer is a professor and head of the Distributed and Parallel Systems Group at the Institute of Computer Science at the University of Innsbruck, Austria. His 32 A BEMaGS F research interests include distributed and parallel systems. Fahringer received a PhD in computer science from the Technical University of Vienna. He is a member of the IEEE and the ACM. Contact him at ____________ tf@dps.uibk.ac.at. Geoffrey Fox is a professor of physics and computer science at Indiana University and a distinguished scientist in its Community Grids Laboratory. His research interests include grids and parallel computing. Fox received a PhD in theoretical physics from Cambridge University. He is a member of the ACM and the IEEE Computer Society and a fellow of the American Physical Society. Contact him at gcf@indiana.edu. ___________ Dennis Gannon is a professor of computer science in the School of Informatics at Indiana University. His research interests include cyberinfrastructure, programming systems and tools, and distributed computing. Gannon received a PhD in mathematics from the University of California, Davis, and a PhD in computer science from the University of Illinois. Contact him at ________________ gannon@cs.indiana.edu. Carole Goble is a professor at the University of Manchester, and director of the myGRID project. Her research interests are the Semantic Web, e-science, and grid communities. Goble received a BSc from Manchester University. She is a member of the IEEE, the ACM, and the British Computer Society. Contact her at _____________________ carole.goble@manchester.ac.uk. Miron Livny is a computer science professor at the University of Wisconsin-Madison. His research interests include high-throughput computing, visual data exploration, and experiment-management environments. Livny received a PhD in computer science from the Weizmann Institute of Science, Israel. Contact him at _____________ miron@cs.wisc.edu. Luc Moreau is a professor of computer science at the University of Southampton. His research interests include largescale open distributed systems and provenance. He received a PhD from the University of Liège, Belgium. He is a fellow of the British Computer Society and a member of the ACM. Contact him at ___________________ L.Moreau@ecs.soton.ac.uk. Jim Myers leads the Cyberenvironments and Technologies Directorate at the National Center for Supercomputing Applications. His research interests include open source collaborative tools. Myers received a PhD in chemistry from the University of California, Berkeley. He is a member of the American Chemical Society, the American Physical Society, the ACM, and the IEEE. Contact him at ________ jimmyers@ ncsa.uiuc.edu. _________ Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page C O V E R A BEMaGS F F E A T U R E The Case for Energy-Proportional Computing Luiz André Barroso and Urs Hölzle Google Energy-proportional designs would enable large energy savings in servers, potentially doubling their efficiency in real-life use. Achieving energy proportionality will require significant improvements in the energy usage profile of every system component, particularly the memory and disk subsystems. E nergy efficiency, a new focus for general-purpose computing, has been a major technology driver in the mobile and embedded areas for some time. Earlier work emphasized extending battery life, but it has since expanded to include peak power reduction because thermal constraints began to limit further CPU performance improvements. Energy management has now become a key issue for servers and data center operations, focusing on the reduction of all energy-related costs, including capital, operating expenses, and environmental impacts. Many energy-saving techniques developed for mobile devices became natural candidates for tackling this new problem space. Although servers clearly provide many parallels to the mobile space, we believe that they require additional energy-efficiency innovations. In current servers, the lowest energy-efficiency region corresponds to their most common operating mode. Addressing this mismatch will require significant rethinking of components and systems. To that end, we propose that energy proportionality should become a primary design goal. Although our experience in the server space motivates these observations, we believe that energy-proportional computing also will benefit other types of computing devices. DOLLARS & CO2 Recent reports1,2 highlight a growing concern with computer-energy consumption and show how current 0018-9162/07/$25.00 © 2007 IEEE Computer trends could make energy a dominant factor in the total cost of ownership.3 Besides the server electricity bill, TCO includes other energy-dependent components such as the cost of energy for the cooling infrastructure and provisioning costs, specifically the data center infrastructure’s cost. To a first-order approximation, both cooling and provisioning costs are proportional to the average energy that servers consume, therefore energy efficiency improvements should benefit all energy-dependent TCO components. Efforts such as the Climate Savers Computing Initiative (www.climatesaverscomputing.org) could help lower worldwide computer energy consumption by promoting widespread adoption of high-efficiency power supplies and encouraging the use of power-savings features already present in users’ equipment. The introduction of more efficient CPUs based on chip multiprocessing has also contributed positively toward more energy-efficient servers.3 However, long-term technology trends invariably indicate that higher performance means increased energy usage. As a result, energy efficiency must improve as fast as computing performance to avoid a significant growth in computers’ energy footprint. SERVERS VERSUS LAPTOPS Many of the low-power techniques developed for mobile devices directly benefit general-purpose servers, including multiple voltage planes, an array of energyefficient circuit techniques, clock gating, and dynamic Published by the IEEE Computer Society 33 December 2007 Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ments because minor traffic fluctuations or any internal disruption, such as hardware or software faults, could tip it over the edge. Moreover, the lack of a 0.025 reasonable amount of slack makes regular operations exceedingly complex because 0.02 any maintenance task has the potential to cause serious service disruptions. Similarly, well-pro0.015 visioned services are unlikely to spend significant amounts of time completely idle because 0.01 doing so would represent a substantial waste of capital. Even during periods of low ser0.005 vice demand, servers are unlikely to be fully idle. Large-scale services usually require hundreds of 0 servers and distribute the load 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 over these machines. In some CPU utilization cases, it might be possible to completely idle a subset of servers Figure 1. Average CPU utilization of more than 5,000 servers during a six-month period. during low-activity periods by, Servers are rarely completely idle and seldom operate near their maximum utilization, for example, shrinking the numinstead operating most of the time at between 10 and 50 percent of their maximum ber of active front ends. Often, utilization levels. though, this is hard to accomplish because data, not just comvoltage-frequency scaling. Mobile devices require high putation, is distributed among machines. For example, performance for short periods while the user awaits a common practice calls for spreading user data across response, followed by relatively long idle intervals of many databases to eliminate the bottleneck that a censeconds or minutes. Many embedded computers, such tral database holding all users poses. as sensor network agents, present a similar bimodal Spreading data across multiple machines improves usage model.4 data availability as well because it reduces the likeliThis kind of activity pattern steers designers to empha- hood that a crash will cause data loss. It can also help size high energy efficiency at peak performance levels hasten recovery from crashes by spreading the recovand in idle mode, supporting inactive low-energy states, ery load across a greater number of nodes, as is done such as sleep or standby, that consume near-zero energy. in the Google File System.6 As a result, all servers must However, the usage model for servers, especially those be available, even during low-load periods. In addition, used in large-scale Internet services, has very different networked servers frequently perform many small backcharacteristics. ground tasks that make it impossible for them to enter Figure 1 shows the distribution of CPU utilization lev- a sleep state. els for thousands of servers during a six-month interWith few windows of complete idleness, servers canval.5 Although the actual shape of the distribution varies not take advantage of the existing inactive energysignificantly across services, two key observations from savings modes that mobile devices otherwise find so Figure 1 can be generalized: Servers are rarely com- effective. Although developers can sometimes restrucpletely idle and seldom operate near their maximum uti- ture applications to create useful idle intervals during lization. Instead, servers operate most of the time at periods of reduced load, in practice this is often difficult between 10 and 50 percent of their maximum utiliza- and even harder to maintain. The Tickless kernel7 exemtion levels. Such behavior is not accidental, but results plifies some of the challenges involved in creating and from observing sound service provisioning and distrib- maintaining idleness. Moreover, the most attractive inacuted systems design principles. tive energy-savings modes tend to be those with the highAn Internet service provisioned such that the average est wake-up penalties, such as disk spin-up time, and load approaches 100 percent will likely have difficulty thus their use complicates application deployment and meeting throughput and latency service-level agree- greatly reduces their practicality. Fraction of time 0.03 34 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page ENERGY EFFICIENCY AT VARYING UTILIZATION LEVELS A BEMaGS F 100 Typical operating region TOWARD ENERGY-PROPORTIONAL MACHINES Addressing the mismatch between the servers’ energy-efficiency characteristics and the behavior of server-class workloads is primarily the responsibility of component and system designers. They should aim to develop machines that consume energy in proportion to the amount of work performed. Such energyproportional machines would ideally consume no power when idle (easy with inactive power modes), nearly no power when very little work is performed (harder), and gradually more power as the activity level increases (also harder). Energy-proportional machines would exhibit a wide dynamic power range—a property that might be rare today in computing equipment but is not unprecedented in other domains. Humans, for example, have an average daily energy consumption approaching that of an old personal computer: about 120 W. However, humans at rest can consume as little as 70 W,8 while being able to sustain peaks of well over 1 kW for tens of minutes, with elite athletes reportedly approaching 2 kW.9 Breaking down server power consumption into its main components can be useful in helping to better CPU contribution to server power Server power usage (percent of peak) Server power consumption responds 90 differently to varying utilization levels. 80 We loosely define utilization as a measure of the application performance— 70 such as requests per second on a Web 60 server—normalized to the perfor50 mance at peak load levels. Figure 2 shows the power usage of a typical 40 energy-efficient server, normalized to 30 its maximum power, as a function of utilization. Essentially, even an energy20 efficient server still consumes about Power 10 Energy efficiency half its full power when doing virtually no work. Servers designed with 0 0 10 20 30 40 50 60 70 80 90 100 less attention to energy efficiency often Utilization (percent) idle at even higher power levels. Seeing the effect this narrow dynamic power range has on such a system’s Figure 2. Server power usage and energy efficiency at varying utilization levels, energy efficiency—represented by the from idle to peak performance. Even an energy-efficient server still consumes red curve in Figure 2—is both enlight- about half its full power when doing virtually no work. ening and discouraging. To derive power efficiency, we simply divide utilization by its cor60 responding power value. We see that peak energy efficiency occurs at peak utilization and drops quickly as 50 utilization decreases. Notably, energy efficiency in the 20 to 30 percent utilization range—the point at which 40 servers spend most of their time—has dropped to less than half the energy efficiency at peak performance. 30 Clearly, such a profile matches poorly with the usage 20 characteristics of server-class applications. 10 0 2005 server (peak) 2007 server (peak) 2007 server (idle) Google servers Figure 3. CPU contribution to total server power for two generations of Google servers at peak performance (the first two bars) and for the later generation at idle (the rightmost bar). understand the key challenges for achieving energy proportionality. Figure 3 shows the fraction of total server power consumed by the CPU in two generations of Google servers built in 2005 and 2007. The CPU no longer dominates platform power at peak usage in modern servers, and since processors are adopting energy-efficiency techniques more aggressively than other system components, we would expect CPUs to contribute an even smaller fraction of peak power in future systems. Comparing the second and third bars in Figure 3 provides useful insights. In the same platform, the 2007 server, the CPU represents an even smaller fraction of total power when the system 35 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Compared to today’s machines, servers with a dynamic power range of Typical operating region 90 percent, shown in Figure 4, could cut 90 by one-half the energy used in data cen80 ter operations.5 They would also lower peak power at the facility level by more 70 than 30 percent, based on simulations 60 of real-world data center workloads. 50 These are dramatic improvements, especially considering that they arise 40 from optimizations that leave peak 30 server power unchanged. The power efficiency curve in Figure 4 fundamen20 tally explains these gains. This server Power 10 Energy efficiency has a power efficiency of more than 80 percent of its peak value for utilizations 0 0 10 20 30 40 50 60 70 80 90 100 of 30 percent and above, with efficiency Utilization (percent) remaining above 50 percent for utilization levels as low as 10 percent. Figure 4. Power usage and energy efficiency in a more energy-proportional server. In addition to its energy-savings This server has a power efficiency of more than 80 percent of its peak value for potential, energy-proportional hardutilizations of 30 percent and above, with efficiency remaining above 50 percent ware could obviate the need for power for utilization levels as low as 10 percent. management software, or at least simplify it substantially, reducing power is idle, suggesting that processors are closer to exhibit- management to managing utilization. Fundamentally, the latency and energy penalties ing the energy-proportional behavior we seek. Two key CPU features are particularly useful for incurred to transition to the active state when starting achieving energy proportionality and are worthy of an operation make an inactive energy-savings mode less useful for servers. For example, a disk drive in a spunimitation by other components. down, deep-sleep state might use almost no energy, but a transition to active mode incurs a latency penalty 1,000 Wide dynamic power range Current desktop and server processors can consume times higher than a regular access latency. Spinning up less than one-third of their peak power at very-low activ- the platters also carriers a large energy penalty. Such a ity modes, creating a dynamic range of more than 70 huge activation penalty restricts spin-down modes to sitpercent of peak power. CPUs targeted at the mobile or uations in which the device will be idle for several minembedded markets can do even better, with idle power utes; this rarely occurs in servers. On the other hand, often reaching one-tenth or less of peak power.10 They inactive energy-savings modes with wake-up penalties achieve this even when not using any performance- of only a small fraction of the regular operations’ latency impacting—or software-visible—energy-saving modes. are more likely to benefit the server space, even if their In our experience, the dynamic power range of all low-energy state operates at relatively higher energy levother components is much narrower: less than 50 per- els than would be possible in deep-sleep modes. cent for DRAM, 25 percent for disk drives, and 15 perActive energy-savings schemes, by contrast, are useful cent for networking switches. even when the latency and energy penalties to transition to a high-performance mode are significant. Since active Active low-power modes modes are operational, systems can remain in lowA processor running at a lower voltage-frequency mode energy states for as long as they remain below certain can still execute instructions without requiring a perfor- load thresholds. Given that periods of low activity are mance-impacting mode transition. It is still active. There more common and longer than periods of full idleness, are no other components in the system with active low- the overheads of transitioning between active energypower modes. Networking equipment rarely offers any savings modes amortize more effectively. low-power modes, and the only low-power modes currently available in mainstream DRAM and disks are fully inactive. That is, using the device requires paying a latency ervers and desktop computers benefit from much and energy penalty for an inactive-to-active mode transiof the energy-efficiency research and development tion. Such penalties can significantly degrade the perforthat was initially driven by mobile devices’ needs. mance of systems idle only at submillisecond time scales. However, unlike mobile devices, which idle for long Server power usage (percent of peak) 100 S 36 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page periods, servers spend most of their time at moderate utilizations of 10 to 50 percent and exhibit poor efficiency at these levels. Energy-proportional computers would enable large additional energy savings, potentially doubling the efficiency of a typical server. Some CPUs already exhibit reasonably energy-proportional profiles, but most other server components do not. We need significant improvements in memory and disk subsystems, as these components are responsible for an increasing fraction of the system energy usage. Developers should make better energy proportionality a primary design objective for future components and systems. To this end, we urge energy-efficiency benchmark developers to report measurements at nonpeak activity levels for a more complete characterization of a system’s energy behavior. ■ Acknowledgments We thank Xiaobo Fan and Wolf-Dietrich Weber for coauthoring the power provisioning study that motivated this work, and Catherine Warner for her comments on the manuscript. References 1. US Environmental Protection Agency, “Report to Congress on Server and Data Center Energy Efficiency: Public Law 109431”; www.energystar.gov/ia/partners/prod_development/ downloads/EPA_Datacenter_Report_Congress_Final1.pdf. ____________________________________ 2. J.G. Koomey, “Estimating Total Power Consumption by Servers in the U.S. and the World”; http://enterprise.amd.com/ Downloads/svrpwrusecompletefinal.pdf. _________________________ 3. L.A. Barroso, “The Price of Performance: An Economic Case for Chip Multiprocessing,” ACM Queue, Sept. 2005, pp. 4853. 4. J. Hill et al., “System Architecture Directions for Networked Sensors,” Proc. SIGOPS Oper. Syst. Rev., ACM Press, vol. 34, no. 5, 2000, pp. 93-104. 5. X. Fan, W.-D. Weber, and L.A. Barroso, “Power Provisioning for a Warehouse-Sized Computer”; http://research.google. com/archive/power_provisioning.pdf. 6. S. Ghemawat, H. Gobioff and S.-T. Leung, “The Google File System”; www.cs.rochester.edu/meetings/sosp2003/papers/ p125-ghemawat.pdf. _____________ 7. S. Siddha, V. Pallipadi, and A. Van De Ven, “Getting Maximum Mileage Out of Tickless,” Proc. 2007 Linux Symp., 2007, pp. 201-208. 8. E. Ravussin et al., “Determinants of 24-Hour Energy Expenditure in Man: Methods and Results Using a Respiratory Chamber”; http://www.pubmedcentral.nih.gov/picrender. fcgi?artid=423919&blobtype=pdf. _____________________ 9. E.F. Coyle, “Improved Muscular Efficiency Displayed as Tour de France Champion Matures”; http://jap.physiology.org/ cgi/reprint/98/6/2191. _____________ A BEMaGS F 10. Z. Chen et al., “A 25W(max) SoC with Dual 2GHz Power Cores and Integrated Memory and I/O Subsystems”; www. ___ pasemi.com/downloads/PA_Semi_ISSCC_2007.pdf. Luiz André Barroso is a distinguished engineer at Google. His interests range from distributed system software infrastructure to the design of Google’s computing platform. Barroso received a PhD in computer engineering from the University of Southern California. Contact him at ____ luiz@ google.com. Urs Hölzle is the senior vice president of operations at Google and a Google Fellow. His interests include largescale clusters, cluster networking, Internet performance, and data center design. Hölzle received a PhD in computer science from Stanford University. Contact him at ____ urs@ google.com. Computer Wants You Computer is always looking for interesting editorial content. In addition to our theme articles, we have other feature sections such as Perspectives, Computing Practices, and Research Features as well as numerous columns to which you can contribute. Check out our author guidelines at www.computer.org/computer/ author.htm ________ for more information about how to contribute to your magazine. 37 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F _____________________________________________________ Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page C O V E R A BEMaGS F F E A T U R E Models and Metrics to Enable Energy-Efficiency Optimizations Suzanne Rivoire, Stanford University Mehul A. Shah and Parthasarathy Ranganathan, Hewlett-Packard Laboratories Christos Kozyrakis, Stanford University Justin Meza, University of California, Los Angeles Power consumption and energy efficiency are important factors in the initial design and day-to-day management of computer systems. Researchers and system designers need benchmarks that characterize energy efficiency to evaluate systems and identify promising new technologies.To predict the effects of new designs and configurations, they also need accurate methods of modeling power consumption. I n recent years, the power consumption of servers and data centers has become a major concern. According to the US Environmental Protection Agency, enterprise power consumption in the US doubled between 2000 and 2006 (www.energystar. gov/ia/partners/prod_development/downloads/EPA_ _______________________________________ Datacenter_Report_Congress_Final1.pdf), and will dou_____________________________ ble again in the next five years. Server power consumption not only directly affects a data center’s electricity costs, but also necessitates the purchase and operation of cooling equipment, which can consume from one-half to one watt for every watt of server power consumption. All of these power-related costs can potentially exceed the cost of purchasing hardware. Moreover, the environmental impact of data center power consumption is receiving increasing attention, as is the effect of escalating power densities on the ability to pack machines into a data center.1 The two major and complementary ways to approach this problem involve building energy efficiency into the initial design of components and systems, and adaptively managing the power consumption of systems or groups of systems in response to changing conditions in the workload or environment. Examples of the former approach include 0018-9162/07/$25.00 © 2007 IEEE Computer • circuit techniques such as disabling the clock signal to a processor’s unused parts; • architectural techniques such as replacing complex uniprocessors with multiple simple cores; and • support for multiple low-power states in processors, memory, and disks. At the system level, the latter approach requires policies to intelligently exploit these low-power states for energy savings. Across multiple systems in a cluster or data center, these policies can involve dynamically adapting workload placement or power provisioning to meet specific energy or thermal goals.1 To facilitate these optimizations, we need metrics to define energy efficiency, which will help designers compare designs and identify promising energy-efficient technologies. We also need models to predict the effects of dynamic power management policies, particularly over many systems. Unlike the significant body of work on power management and optimization, there has been relatively little focus on metrics and models. We address the challenges in defining metrics for energy efficiency with a specific case study on JouleSort, which provides a complete, full-system benchmark for energy efficiency across a variety of system classes.2 The “Approaches to Power Modeling in Computer Systems” Published by the IEEE Computer Society 39 December 2007 Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Approaches to Power Modeling in Computer Systems Power models used in simulators of proposed hardware trade speed and portability for increased accuracy, relying on detailed knowledge of component architecture and circuit technology. Wattch1 is a widely used CPU power model that estimates the power costs of accessing different parts of the processor and combines this information with activity counts from a performance simulator to yield power estimates. Similar models have been proposed for other components, including memory, disks, and networking, as well as complete systems.2,3 These simulators are highly accurate, but also closely tied to specific systems and simulation infrastructures that are much slower than actual hardware. Models used in online powermanagement policies, for which 8 speed is a first-class constraint, cannot rely on such detailed simu7 lation. Using real-time system 6 events instead of simulated activity counts addresses this drawback. 5 Frank Bellosa proposed using ppred, t = 14.45 + 23.61 * ucpu, t – 0.85 * umem, t + 22.32 * udisk, t + 0.03 * unet, t 4 processor performance counter registers to provide on-the-fly 3 power characterization of real systems.4 His simple and portable 2 model used the counts of instruc1 tions executed and memory accesses to drive the selection of a 0 Matrix Stream SPECjbb SPECweb SPECint SPECfp processor’s frequency states. More Benchmark detailed and processor-specific performance-counter-based models have been developed to model Figure A. Accuracy of the model created by Mantis9 for a low-power blade.The equation both power5 and thermal6 propershows the power predicted at a given time t as a function of CPU utilization, number of ties. Finally, since researchers develmemory and disk accesses, and number of network accesses sampled at that time. Each utioped performance counter options lization input u is given as a percentage of its maximum value.The average error of this with application profiling rather coarse-grained linear model is less than 10 percent for every benchmark—sufficient for than power estimation in mind, most scheduling optimizations. Average prediction error (percent) Power models are fundamental to energy-efficiency research, whether the goal is to improve the components’ and systems’ design or to efficiently use existing hardware. Developers use these models offline to evaluate proposed designs, and online in policies to exploit component power modes within a system, or to efficiently distribute work across several systems. An ideal power model has several properties. First, it must be accurate. It should also be portable across a variety of existing and future hardware designs, and applicable under a variety of workloads. Finally, it should be cost-effective in its hardware and software requirements and execute swiftly. sidebar describes different approaches to modeling power consumption in components, systems, and data centers. ENERGY-EFFICIENCY METRICS An ideal benchmark for energy efficiency would consist of a universally relevant workload, a metric that balances power and performance in a universally appropriate way, and rules that provide impossible-to-circumvent, fair comparisons. Since this benchmark is impossible, several different approaches have addressed pieces of the energyefficiency evaluation problem, from chips to data centers. Table 1 summarizes these approaches. Each metric in Table 1 addresses a particular energyrelated problem, from minimizing power consumption 40 in embedded processors (EnergyBench) to evaluating the efficiency of data center cooling and power provisioning (Green Grid metrics). However, only JouleSort specifies a workload, a metric to compare two systems, and rules for running the benchmark. At the processor level, Ricardo Gonzalez and Mark Horowitz argued in 19963 that the energy-delay product provides the appropriate metric for comparing two designs. They observed that a chip’s performance and power consumption are both related to the clock frequency, with performance directly proportional to, and power consumption increasing quadratically with, clock frequency. Comparing processors based on energy, which is the product of execution time and Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Ismail Kadayif and colleagues proposed an interface based on “energy counters” that would virtualize the existing performance counters.7 Processor performance counters can be used to estimate processor and memory power consumption, but do not take other parts of the system, such as I/O, into account. Some optimizations, such as data-center-level optimizations that turn off unused machines, must consider the full-system power. In this case, OS utilization metrics can be used to model the base system components quickly, portably, and with reasonable accuracy. Taliver Heath and colleagues8 and Dimitris Economou and colleagues9 build linear models based on OS-reported utilization of each component. Both approaches require an initial calibration phase, in which developers connect the system to a power meter and run microbenchmarks to stress each component. They then fit the utilization data to the power measurements to construct a model. Figure A shows an example of one such model.9 Parthasarathy Ranganathan and Phil Leech used a similar approach to predict both power and performance by constructing lookup tables based on utilization.10 Finally, researchers from Google found that an even simpler model, based solely on OS-reported processor utilization, proved sufficiently accurate to enable optimizations over a large group of homogeneous machines.11 Optimizations for energy efficiency rely on accurate, fast, cost-effective, and portable power models. While many models have been developed to address individual needs, creating systematic methods of generating widely portable and highly accurate models remains an open problem. Such methods could facilitate further innovations in energy-efficient system design and management. References A BEMaGS F work for Architectural-Level Power Analysis and Optimizations,” Proc. Int’l Symp. Computer Architecture (ISCA), ACM Press, 2000, pp. 83-94. 2. W. Ye et al., “The Design and Use of SimplePower: A CycleAccurate Energy Estimation Tool,” Proc. Design Automation Conf. (DAC), IEEE CS Press, 2000, pp. 340-345. 3. S. Gurumurthi et al., “Using Complete Machine Simulation for Software Power Estimation: The SoftWatt Approach,” Proc. High-Performance Computer Architecture Symp. (HPCA), IEEE CS Press, 2002, pp. 141-150. 4. F. Bellosa, “The Benefits of Event-Driven Energy Accounting in Power-Sensitive Systems,” Proc. SIGOPS European Workshop, ACM Press, 2000, pp. 37-42. 5. G. Contreras and M. Martonosi, “Power Prediction for Intel XScale Processors Using Performance Monitoring Unit Events,” Proc. Int’l Symp. Low-Power Electronics and Design (ISLPED), ACM Press, 2005, pp. 221-226. 6. K. Skadron et al., “Temperature-Aware Microarchitecture: Modeling and Implementation,” ACM Trans. Architecture and Code Optimization, vol. 1, no. 1, 2004, pp. 94-125. 7. I. Kadayif et al., “vEC: Virtual Energy Counters,” Proc. PASTE, ACM Press, 2001, pp. 28-31. 8. T. Heath et al., “Energy Conservation in Heterogeneous Server Clusters,” Proc. ACM SIGPLAN Symp. Principles and Practice of Parallel Programming (PpoPP), ACM Press, 2005, pp. 186-195. 9. D. Economou et al., “Full-System Power Analysis and Modeling for Server Environments”; http://csl.stanford.edu/ %7Echristos/publications/2006.mantis.mobs.pdf. ______________________________ 10. P. Ranganathan and P. Leech, “Simulating Complex Enterprise Workloads Using Utilization Traces”; www.hpl.hp.com/ personal/Partha_Ranganathan/papers/2007/2007_caecw_ ____________________________________ bladesim.pdf. ________ 11. X. Fan, W-D. Weber, and L.A. Barroso, “Power Provisioning for a Warehouse-Sized Computer,” Proc. Int’l Symp. Computer Architecture (ISCA), ACM Press, 2007, pp. 13-23. 1. D. Brooks, V. Tiwari, and M. Martonosi, “Wattch: A Frame- power, would therefore motivate processor designers to focus solely on lowering clock frequency at the expense of performance. On the other hand, the energy-delay product, which weighs power against the square of execution time, would show the underlying design’s energy efficiency rather than merely reflecting the clock frequency. In the embedded domain, the Embedded Microprocessor Benchmark Consortium (EEMBC) has proposed the EnergyBench4 processor benchmarks. EnergyBench provides a standardized data acquisition infrastructure for measuring processor power when running one of EEMBC’s existing performance benchmarks. Benchmark scores are then reported as “netmarks per Joule” for networking benchmarks and “telemarks per Joule” for telecommunications benchmarks. The single-system level is the target of several recent metrics and benchmarking efforts. Performance per watt became a popular metric for servers once power became an important design consideration. Performance is typically specified with either MIPS or the rating from peak performance benchmarks like SPECint or TPC-C. Sun Microsystems has proposed the SWaP (space, watts, and performance) metric to include data center space efficiency as well as power consumption.5 Two evolving standards in system-level energy efficiency are the US government’s Energy Star certification guidelines for computers and the SPEC Power and Performance 41 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Table 1. Summary of energy efficiency benchmarks and metrics. Benchmark Metric Level Domain Workload Comment Analysis tool PerformanceN per watt Any Any Unspecified Different balances of performance and power are important in different contexts. N = 0 represents power alone, and N = 2 corresponds to the energydelay product. EnergyBench SWaP Throughput per Joule Performance/(space watts) Processor System(s) Embedded Enterprise EEMBC benchmarks Unspecified Energy Star certification: workstations Certify if “typical” power is less than 35 percent of “maximum” power System Enterprise Energy Star certification: other systems SPEC Power and Performance JouleSort Certify if each mode is below a predefined threshold for that system class Not yet released System System Mobile, desktop, small server Enterprise Sleep, idle, and standby power (typical); Linpack and SPECviewperf (maximum) Sleep, idle, and standby modes Records sorted per Joule System Green Grid DCE Percent of facility power that Data center reaches IT equipment Work done/total facility power (W) Data center Green Grid DCPE committee’s upcoming benchmark. Energy Star is a designation given by the US government to highly energyefficient household products, and has recently been expanded to include computers.6 For most system classes, systems with idle, sleep, and standby power consumptions below a certain threshold will receive the Energy Star rating. For workstations, however, the Energy Star rating requires that the “typical” power—a weighted function of the idle, sleep, and standby power consumptions—not exceed 35 percent of the “maximum power” (the power consumed during the Linpack and SPECviewperf benchmarks, plus a factor based on the number of installed hard disks). Energy Star certification also requires that a system’s power supply efficiency exceed 80 percent. The SPEC power and performance benchmark remains under development.7 The workload will be server-side Java-based, and designed to exercise the system at a variety of usage levels, since servers tend to be underutilized in data center environments. The committee expects to release the specific workload and metric of comparison in late 2007. At the data center level, metrics have been proposed to guide holistic optimizations. To optimize data center cooling, Chandrakant Patel and others have advocated a metric based on weighing performance against the 42 Mobile, desktop, enterprise Enterprise Enterprise Server-side Java under varying loads External sort Addresses both space and power concerns Expected late 2007 Has three benchmark classes with different workload size n/a Not yet determined exergy destroyed. Roughly speaking, exergy is the energy available for doing useful work.8 Finally, the Green Grid, an industrial consortium that includes most major hardware vendors, recently introduced the data center efficiency metric.9 The Green Grid proposal defines DCE as the percentage of total facility power that goes to the “IT equipment”—primarily compute, storage, and network. In the long term, rather than using IT equipment power as a proxy for performance, the Green Grid advocates data center performance efficiency, or the useful work divided by the total facility power. Each of these metrics is useful in evaluating energy efficiency in a particular context, from embedded processors to underutilized servers to entire data centers. However, researchers have not methodically addressed energy-efficiency metrics for many important computing domains. For example, there are no full-system benchmarks that specify a workload, a metric to compare two systems, and rules for running the benchmark. The recently proposed JouleSort benchmark addresses this space. JOULESORT BENCHMARK We designed the JouleSort benchmark with several goals in mind. First, the benchmark should evaluate the Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F power-performance tradeoff—that is, the Table 2. Summary of sort benchmarks.10 benchmark score should not reward high performance or low power alone. Two reaBenchmark Description Status sonable metrics for the benchmark are thus energy (the product of average power conPennySort Sort as many records as possible for one Active sumption and execution time) and the cent, assuming a 3-year depreciation. energy-delay product, which places more MinuteSort Sort as many records as possible in less Active emphasis on performance. than a minute. We chose energy for two reasons. First, plenty TerabyteSort Sort a Tbyte of data (10 billion records) as Active of performance benchmarks already exist, so quickly as possible. we wanted to be sure our benchmark emphaDatamation Sort 1 million records as quickly as Deprecated sized power. Second, the tradeoff between perpossible. formance and power at the system level does JouleSort Sort a fixed number of records (approx. Proposed not display the straightforward quadratic rela10 Gbytes, 100 Gbytes, 1 Tbyte) using as tionship seen at the processor level, which little energy as possible. motivated use of the energy-delay metric. Further, the benchmark should evaluate a system’s peak energy efficiency, which for today’s sys- records a system can sort for one penny, assuming a tems occurs at peak utilization. While peak utilization three-year depreciation. MinuteSort and TerabyteSort offers a realistic scenario in some domains, data center measure a system’s pure performance in sorting for a servers in particular are notoriously underutilized. fixed time of one minute and a fixed data set of one However, benchmarking at peak utilization is justified Tbyte, respectively. JouleSort, to measure the powerfor several reasons. First, peak utilization is simpler to performance tradeoff, is thus a logical addition to the define and measure, and it makes the benchmark more sort benchmark repertoire. The original Datamation sort difficult to circumvent. Additionally, knowing the upper benchmark compared the amount of time systems took bound on energy efficiency for a particular system is use- to sort 1 million records; it is now deprecated since this ful. In enterprise environments, for example, this upper task is trivial on modern systems. bound provides a target for server consolidation. The workload can be summarized as follows: Sort a Next, the benchmark should be balanced. It should file consisting of randomly permuted 100-byte records stress all core system components, and the metric should with 10-byte keys. The input file must be read from— incorporate the energy that all components use. It should and the output file written to—nonvolatile storage. The also be representative of important workloads and sim- output file must be newly created rather than overwritple to implement and administer. ing the input file, and all intermediate files that the sort Finally, the benchmark should be inclusive, encom- program uses must be deleted. passing as many past, current, and future systems as posThis workload meets our benchmark goals satisfacsible. For inclusiveness, the benchmark must be torily. It is balanced, stressing I/O, memory, the CPU, meaningful and measurable on as many system classes the OS, and the file system. It is representative and as possible. The workload and metric should apply to a inclusive; it resembles sequential, I/O-intensive datamanagement workloads that are found on most platwide range of technologies. forms, from cell phones processing multimedia data to Benchmark workload clusters performing large-scale parallel data analysis. For our benchmark’s workload, we chose to use the The Sort Benchmark’s longevity testifies to its enduring external sort from the sort benchmarks’ specification applicability as technology changes. (http://research.microsoft.com/research/barc/SortBench mark/default.htm). Benchmark metric _____________ External sort has been a benchmark of interest in the database community since 1985, and Designing a metric that allows fair comparisons across researchers have used it to understand the system-level systems and avoids loopholes that obviate the benchmark effectiveness of algorithmic and component improve- presents a major challenge in benchmark development. ments and identify promising technology trends. For JouleSort, we seek to evaluate the power-performance Previous sort benchmark winners have foreshadowed balance of different systems, giving power and perforthe transition from supercomputers to commodity clus- mance equal weight. We could have defined the JouleSort ters, and recently showed the promise of general-pur- benchmark score in three different ways: pose computation on graphics processing units (GPUs).10 The sort benchmarks currently have three active cat• Set a fixed energy budget for the sort, and compare egories, as summarized in Table 2. PennySort is a pricesystems based on the number of records sorted performance benchmark that measures the number of within that budget. 43 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F The first problem, then, is the disincentive to continue sorting beyond the largest one-pass sort. With a budget of one 16,000 minute, this particular machine would 14,000 achieve its best records-sorted-per-Joule rating if it sorted 15 million records, 12,000 which takes 10 seconds, and went into a 10,000 low-power sleep mode for the remaining 50 seconds. In the extreme case, a system 8,000 optimized for this benchmark could spend 6,000 most of the benchmark’s duration in sleep mode—thus voiding the goal of measur4,000 ing a utilized system’s efficiency. 2,000 The second problem is the (N lg N) 0 algorithmic complexity of sort, which 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10 causes the downward trend in efficiency Records sorted for large data sets. While constant factors initially obscure this complexity, once the Figure 1. Problems with using a fixed time budget and a metric of records sorted sort becomes CPU-bound, the number of per Joule.The dramatic drop in efficiency at the transition from one-pass to records sorted per Joule begins to decrease two-pass sorts (here, at 15 million records) creates an incentive to sleep for because the execution time now increases some, or even most, of the time budget.The (N lg N) complexity of sort causes superlinearly with the number of records. the slow drop-off in efficiency for large data sets at the rightmost part of the In light of these problems with a fixed time graph and creates a similar problem. budget and fixed energy budget, we settled on using a fixed input size. This deci• Set a fixed time budget for the sort, and compare sys- sion necessitates multiple benchmark classes, similar to tems based on the number of records sorted and the the TPC-H benchmark, since different workload sizes amount of energy consumed, expressed as records are appropriate to different system classes. The JouleSort classes are 100 million records (about 10 Gbytes), 1 bilsorted per Joule. • Set a fixed workload size for the sort, and compare lion records (about 100 Gbytes), and 10 billion records systems based on the amount of energy consumed. (about 1 Tbyte). The metric of comparison then becomes the minimum energy or records sorted per Joule, which The fixed-energy budget and fixed workload both are equivalent for a fixed workload size. We prefer the latter metric because it highlights effihave the drawback that a single fixed budget will not be applicable to all classes of systems, necessitating multi- ciency more clearly and allows rough comparisons across ple benchmark classes and updates to the class defini- different benchmark classes, with the caveats we have tions as technology changes. The fixed-energy budget described. We do anticipate that the benchmark classes has the further drawback of being difficult to bench- will change as systems become more capable. However, mark. Since energy is the product of power and time, it since sort performance is improving more slowly than is affected by variations in both quantities. Measurement Moore’s law, we expect the current classes to be relevant error from power meters only compounds this problem. for at least five years. Therefore, given our criteria, the By contrast, using a reasonably low fixed-time bud- fixed input size offers the most reasonable option. get and a metric of records sorted per Joule would avoid this problem; however, two more serious issues elimi- Energy measurement nate it from consideration. Figure 1 illustrates these, While we can borrow many of the benchmark rules showing the records sorted per Joule for our best-per- from the existing sort benchmarks, energy measurement forming system while running different workload sizes. requires additional guidelines. The most important areas From the left, the smallest data set sizes take only a few to consider are the boundaries of the system to be measeconds and thus poorly amortize the startup overhead. sured, constraints on the ambient environment, and As data sets grow larger, this overhead amortizes better, acceptable methods of measuring power consumption. while efficiency increases, up to 15 million records. This is The energy consumed to power the physical system the largest data set that fits completely in memory. For executing the sort is measured from the wall outlet. larger sizes, the system must temporarily write data to disk, This approach accounts for power supply inefficiendoubling the amount of I/O and decreasing performance cies in converting from AC to DC power, which can be dramatically. After this transition, energy efficiency stays significant.1 If a component remains unused in the sort relatively constant, with a slow trend downward. and cannot be physically removed from the system, JouleSort score (records/J) 18,000 44 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Energy efficiency of previous sort benchmark winners First, to understand historical trends in energy efficiency, we retrospectively applied our benchmark to previous sort benchmark winners over the past decade, computing their scores in records sorted per Joule. Since there are no power measurements for these systems, we estimated the power consumption based on the benchmark winners’ posted reports on the Sort Benchmark Web site, which include both hardware configuration information and performance data. The estimation methodology relies on the fact that these historical winners have used desktop- and server-class components that should be running at or near peak power for most of the sort. Therefore, we JouleSort score (records/J) we include its power consumption in the Table 3. Summary of JouleSort benchmark definitions. measurement. The benchmark accounts for the energy Workload External sort consumed by elements of the cooling infrastructure, such as fans, that physically conBenchmark classes 108 records (10 Gbytes), 109 records (100 Gbytes), nect to the hardware. While air conditioners, 1010 records (1 Tbyte) blowers, and other cooling devices consume Benchmark categories Daytona = commercially supported hardware and software significant amounts of energy in data cenIndy = “no holds barred” implementations ters, it would be unreasonable to include Metric Energy to sort a fixed number of records (records sorted them for all but the largest sorting systems. per Joule) We do specify that the ambient temperature Energy measurement Measure power at the wall, subject to EPA power meter at the system’s inlets be maintained at guidelines between 20° to 25° C—typical for data cenEnvironment Maintain ambient temperature of 20°-25° C ter environments. Finally, energy consumption should be measured as the product of the wall clock time used for the can approximate component power consumption as sort and the average power over the sort’s execution. The constant over the sort’s length. execution time will be measured as the existing sort We validated our estimation methodology on singlebenchmarks specify. The easiest way to measure the node desktop- and server-class systems, for which power is to plug the system into a digital power meter, the estimates were accurate within 5 to 25 percent— which then plugs into the wall; the SPEC Power com- sufficiently accurate to draw high-level conclusions. mittee and the Energy Star guidelines have jointly proThe historical data, shown in Figure 2, supports a few posed minimum power meter requirements,6,7 which we observations. First, the PennySort winners tend to be the most adopt for JouleSort as well. Finally, we define two benchmark categories: Daytona, for commercially supported energy-efficient systems, for the simple reason that hardware and software, and Indy, which is unconstrained. PennySort is the only benchmark to weigh performance against a resource constraint. While low cost and low Table 3 summarizes the final benchmark definition. power consumption do not always correlate, both metrics tend to encourage minimizing the number of JOULESORT BENCHMARK RESULTS Using this benchmark, we evaluated energy efficiency for a variety of computer 3,500 systems. We first estimated the energy efficiency of previous Sort Benchmark winPennySort Daytona 3,000 PennySort Indy ners and then experimentally evaluated MinuteSort Daytona different systems with the JouleSort MinuteSort Indy 2,500 benchmark. Terabyte Daytona 2,000 Terabyte Indy Datamation 1,500 1,000 500 0 1996 1998 2000 2002 2004 2006 2008 Sort benchmmark winners by year Figure 2. Estimated energy efficiency, in records sorted per Joule, of historical Sort Benchmark winners.The Daytona category is for commercially supported sorts, while the Indy category has no such restrictions.The pink arrow shows the energy efficiency trend for cost-efficient sorts, which is improving at a rate of 25 percent per year.The blue arrow shows the trend for performance-oriented sorts, whose energy efficiency is improving at 13 percent per year. Both of these rates fall well below the rates of improvement in performance and cost performance. 45 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F power measures the entire enclosure, which is designed to deliver power to 15 blades and is thus both overprovisioned Name Description and inefficient at this low load. We thereLaptop A modern laptop with an Intel Core 2 Duo processor and 3 Gbytes fore include both the wall power of the of RAM enclosure—blade-wall—and a more realBlade-wall A single low-power blade plus the full wall power of its enclosure istic calculation of the power consump(designed for 16 blades) tion for the blade itself, plus a Blade-amortized A single low-power blade plus its proportionate share of the proportionate share of the enclosure enclosure power overhead, which we call blade-amortized. Standard server A standard server with Intel Xeon processor, 2 Gbytes of RAM, The laptop and the fileserver proved to and 2 hard disks be the two most efficient “off-the-shelf” Fileserver A fileserver with 2 disk trays containing 6 disks per tray systems by far—both have energy effiCoolSort A desktop with a high-end mobile processor, 2 Gbytes of RAM, ciency similar to the most efficient hisand 13 SATA laptop disks torical system. The file server’s high Gumstix An ultra-low-power system used in embedded devices energy efficiency is not surprising because Soekris A board typically used for networking applications the CPU and I/O both operate at peak VIA-laptop A VIA picoITX multimedia machine with laptop hard disks utilization, which corresponds to peak VIA-flash A VIA picoITX multimedia machine with flash drives energy efficiency for today’s equipment. The laptop, however, shows high energy efficiency even though its CPU is drasticomponents and using lower-performance components cally underutilized. These results suggest that a benchwithin a class. mark-winning JouleSort machine could be constructed Second, comparing this graph to the published per- by creating a balanced sorting machine out of mobileformance records shows that the energy-efficiency scores class components. of sort benchmark winners have not improved at nearly Based on these insights, we identified two approaches the same rate as performance or price-performance to custom-assembled machines. The first builds a scores. The PennySort winners have improved in both machine from mobile-class components and attempts performance and cost efficiency at rates greater than 50 to maximize performance. The second tries to minimize percent per year. Their energy efficiency, on the other power while still designing a machine with reasonable hand, has improved by just 25 percent per year, most of performance. Both approaches lead to energy efficiencies which came in the past two years. more than 2.5 times greater than in previous systems. The winners of the performance sorts (MinuteSort, The former approach led to the design of the CoolSort TerabyteSort, and Datamation) have improved their machine. CoolSort uses a high-end mobile CPU conperformance by 38 percent per year, but have improved nected to 13 SATA laptop disks over two PCI-Express energy efficiency by only 13 percent per year. It remains interfaces. The laptop disks use less than one-fifth of the unclear whether these sort benchmark contest winners power of server-class disks, while providing about onewere the most energy-efficient systems of their time, half the bandwidth. At 13 disks, the CPU is fully utilized which suggests the need for a benchmark to track energy during the input pass of the sort, and the motherboard efficiency trends. and disk controllers cannot provide any additional I/O bandwidth. In the 10-Gbyte and 100-Gbyte categories, Current-system and custom-configuration CoolSort’s scores of approximately 11,500 records sorted per Joule are more than three times better than energy efficiency We ran the JouleSort benchmark on a variety of sys- those of any previously measured or estimated systems. tems, including off-the-shelf machines representing Since these notebook-class components have proven major system classes, as well as specialized sorting sys- more energy efficient than their desktop- or server-class tems we created from commodity components. Table 4 counterparts, it makes sense to ask whether a system with summarizes these systems. Since we focus chiefly on even lower-power components could be more energy comparing hardware configurations, we use Ordinal efficient than CoolSort. We examined three embeddedTechnologies’ NSort software for all our experiments. class systems: a Gumstix device; a Soekris machine, typThe commodity machines span several classes of sys- ically found in routers and networking equipment; and tems: a laptop, low-power blade, standard server, and a Via picoITX-based machine, typically used for embedfileserver. In all systems but the fileserver, the CPU is ded multimedia applications. Figure 3 shows the underutilized in the sort because I/O is the bottleneck; JouleSort results of all our measured systems. CPU and I/O utilizations balance in the file server. We Vendors use the smallest and lowest power of these give two measurements for the blade because the wall systems, the Gumstix, in a variety of embedded devices. Table 4. Systems for which the JouleSort rating was experimentally measured. 46 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F JouleSort score (records/J) Our version uses a 600-MHz ARM 18,000 processor, 128 Mbytes of memory, and an 8-Gbyte CompactFlash card. Its power 16,000 consumption is a mere 2 W, but the I/O 14,000 bandwidth is low; a 100-Mbyte in-memory sort takes 137 seconds, sorting 3,650 12,000 records per Joule at a bandwidth of 730 10,000 Via-laptop Via-flash Kbps. For a 2-pass, 1-Gbyte sort, the CoolSort Soekris energy efficiency would probably drop to 8,000 Gumstix Laptop about 1,820 records per Joule. Fileserver Blade-wall 6,000 Blade-amortized Std. server Moving up in power and performance, the Soekris board we used, designed for 4,000 networking equipment, contains a 2662,000 MHz AMD Geode processor, 256 Mbytes of memory, and an 8-Gbyte CompactFlash 0 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10 1.0E+11 card. The power used during sort is 6 W, Records sorted three times that of the Gumstix, but the sort bandwidth is a much higher 3.7 Mbps, yielding 5,945 records sorted per Joule for Figure 3. Measured JouleSort scores of commodity and custom machines.The a 1-Gbyte sort. We determined that the I/O commodity machines, marked with dashes, performed less well than the interface caused the bottleneck, not the custom systems. processor or I/O device itself. The final system, the Via picoITX, has a 1-GHz flash storage over traditional disks. GPUTeraSort’s sucprocessor and 1 Gbyte of DDR2 memory. We tried both cess among the historical sort benchmark winners shows laptop disks and flash as I/O devices; the flash used less that the high performance of GPUs comes at a relatively power and provided higher bandwidth. For a 1-Gbyte small energy cost, although it is unclear whether this will sort, the flash-based version consumed 15 W and sorted continue to hold as GPUs grow ever more power hungry. 10,548 records per Joule—a number close to that of Finally, because sort is a highly parallelizable algorithm, we CoolSort’s two-pass sorts. Although laptop disks are speculate that multicores will be excellent processors in theoretically faster, the limitations of the board allowed energy-efficient sorting systems. Although JouleSort addresses a computer system’s for fewer laptop disks than flash disks, and thus the flash energy efficiency, energy is just one piece of the system’s configuration gave more total I/O bandwidth. The CoolSort and VIA machines improve upon the total cost of ownership (TCO). From the system purprevious year’s efficiency by more than 250 percent, a chaser’s viewpoint, a TCO-Sort would be the most desirmarked departure from the 12 to 25 percent yearly able sort benchmark; however, the TCO components improvement rates over the past decade. Thus, creating vary widely from user to user. Combining JouleSort and benchmarks helps to recognize and drive improvements PennySort to benchmark the costs of purchasing and powering a system is a possible first step. For the highin energy efficiency. efficiency machines we studied, the VIA achieves a JouleSort score comparable to CoolSort, at a much INSIGHTS AND FUTURE WORK The highest-scoring JouleSort machines provide sev- lower price: $1,158 versus $3,032. This result highlights eral insights into system design for energy efficiency. In the potential of flash as a cost-effective technology for the CoolSort machine, we chose components for their achieving high energy efficiency. An emerging area of concern is the scaledown effipower-performance tradeoffs and connected them with high-performance interfaces. While CoolSort’s mobile ciency of components and systems—that is, their ability processor combined with 13 laptop disks offers an to reduce power consumption in response to low utiextreme example, it does highlight the promise of lization.11 Traditionally, components have consumed designing reasonably well-performing servers from well over half their peak power, even when underutimobile components. On the other hand, the lower- lized or idle. Manufacturers are starting to address this power machines suffered because of the limited perfor- inefficiency. JouleSort captures scaledown efficiency to mance of their I/O interfaces, rather than the flash a small extent, since CPU and I/O will not be perfectly devices or CPU. Integration of flash memory closer to balanced during both sort phases, but it does not necthe CPU could create more energy-efficient systems. essarily assess efficiency at low utilization. Second, JouleSort continues the Sort Benchmark’s traFinally, JouleSort can be extended to include metrics of dition of identifying promising new technologies. The VIA importance in the data center. The benchmark’s current system demonstrates the energy efficiency advantages of version does not account for losses in power delivery at 47 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page the data center or rack level. An appropriate benchmark in that setting might be an Exergy JouleSort where the metric of interest is records sorted per Joule of exergy8 expended. A s concerns about enterprise power consumption continue to increase, we need metrics that assess and improve energy efficiency. The JouleSort benchmark can help assess improvements in end-to-end, system-level energy efficiency. This addition to the family of sort benchmarks provides a simple and holistic way to chart trends and identify promising new technologies. The most energy-efficient sorting systems use a variety of emerging technologies, including low-power mobile components, and flash-based storage. ■ A BEMaGS F 7. Standard Performance Evaluation Corporation (SPEC), “SPEC Power and Performance Committee”; www.spec.org/ specpower. ______ 8. C.D. Patel, “A Vision of Energy-Aware Computing from Chips to Data Centers,” Proc. ISMME, Japan Soc. Mechanical Engineers, 2003, pp. 141-150. 9. The Green Grid, “Green Grid Metrics: Describing Datacenter Power Efficiency,” 2007; www.thegreengrid.org/gg_ _________________ content/Green_Grid_Metrics_WP.pdf. _______________________ 10. N.K. Govindaraju et al., “GPUTeraSort: High Performance Graphics Coprocessor Sorting for Large Database Management,” Proc. SIGMOD Conf., ACM Press, 2006, pp. 325336. 11. R. Mayo and P. Ranganathan, “Energy Consumption in Mobile Devices: Why Future Systems Need RequirementsAware Energy Scale-Down,” Special Issue on Power Management, LNCS, Springer, 2003, pp. 26-40. References 1. C.D. Patel and P. Ranganathan, “Enterprise Power and Cooling: A Chip-to-Datacenter Perspective,” tutorial, Proc. Int’l Conf. Architectural Support for Programming Languages and Operating Systems (ASPLOS 2006), ACM Press, 2006, pp.1-101. 2. S. Rivoire et al., “JouleSort: A Balanced Energy-Efficiency Benchmark,” Proc. SIGMOD Conf., ACM Press, 2007, pp. 365-376. 3. R. Gonzalez and M. Horowitz, “Energy Dissipation in General-Purpose Microprocessors,” IEEE J. Solid-State Circuits, Sept. 1996, pp. 1277-1284. 4. Embedded Microprocessor Benchmark Consortium (EEMBC), “EnergyBench v 1.0 Power/Energy Benchmarks”; www.eembc. org/benchmark/power_sl.asp. 5. Sun Microsystems, “SWaP (Space, Watts, and Performance) Metric”; www.sun.com/servers/coolthreads/swap. 6. US Environmental Protection Agency, “ENERGY STAR Program Requirements for Computers: Version 4.0”; www.energystar.gov/ia/partners/prod_development/revisions/ downloads/computer/Computer_Spec_Final.pdf. _____________________________ Submit your manuscript online! Visit http://computer.org/computer and click on “Write for Computer”. IEEE Computer Society 48 Suzanne Rivoire is a PhD candidate in the Department of Electrical Engineering, Stanford University. Her research interests include energy-efficient system design, data center power management, and data-parallel architectures. Rivoire received an MS in electrical engineering from Stanford University. Contact her at ______________ rivoire@stanford.edu. Mehul A. Shah is a research scientist in the Storage Systems Department at Hewlett-Packard Laboratories. His research interests include energy efficiency of computer systems, database systems, distributed systems, and long-term digital preservation. Shah received a PhD in computer science from the University of California, Berkeley. Contact him at mehul.shah@hp.com. ______________ Parthasarathy Ranganathan is a principal research scientist at HP Laboratories. His research interests include lowpower design, system architecture, and parallel computing. Ranganathan received a PhD in electrical and computer engineering from Rice University. Contact him at _____ partha. ranganathan@hp.com. _______________ Christos Kozyrakis is an assistant professor of electrical engineering and computer science at Stanford University. His research interests include transactional memory, architectural support for security, and power management techniques. Kozyrakis received a PhD in computer science from the University of California, Berkeley. Contact him at ____ christos@ee.stanford.edu. ______________ Justin Meza is an undergraduate in computer science at the University of California, Los Angeles. His research interests include open source software, Web standards, electronics, and programming. Contact him at _________ justin.meza@ ucla.edu. ______ Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Welcomes Your Contribution Computer magazine looks ahead to future technologies • Computer, the flagship publication of the IEEE Computer Society, publishes peer-reviewed technical content that covers all aspects of computer science, computer engineering, technology, and applications. • Articles selected for publication in Computer are edited to enhance readability for the nearly 100,000 computing professionals who receive this monthly magazine. • Readers depend on Computer to provide current, unbiased, thoroughly researched information on the newest directions in computing technology. To submit a manuscript for peer-review, see Computer’s author guidelines: www.computer.org/computer/ author.htm ______________ Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page C O V E R A BEMaGS F F E A T U R E The Green500 List: Encouraging Sustainable Supercomputing Wu-chun Feng and Kirk W. Cameron Virginia Tech The performance-at-any-cost design mentality ignores supercomputers’ excessive power consumption and need for heat dissipation and will ultimately limit their performance. Without fundamental change in the design of supercomputing systems, the performance advances common over the past two decades won’t continue. A lthough there’s now been a 10,000-fold increase since 1992 in the performance of supercomputers running parallel scientific applications, performance per watt has only improved 300-fold and performance per square foot only 65-fold. In response to the lagging power and space-efficiency improvements, researchers have had to design and construct new machine rooms, and in some cases, entirely new buildings. Compute nodes’ exponentially increasing power requirements are a primary driver behind this less efficient use of power and space. In fact, the top supercomputers’ peak power consumption has been on the rise over the past 15 years, as Figure 1 shows. Today, several of the most powerful supercomputers on the TOP500 List (www.top500.org) require up to 10 megawatts of peak power—enough to sustain a city of 40,000. And even though IBM Blue Gene/L, the world’s fastest machine, was custom-built with low-power components, the system still consumes several megawatts of power. At anywhere from $200,000 to $1.2 million per megawatt, per year, these are hardly low-cost machines. THE ENERGY CRISIS IN SUPERCOMPUTING Power is a disruptive technology that requires us to rethink supercomputer design. As a supercomputer’s nodes consume and dissipate more power, they must be spaced out and aggressively cooled. Without exotic cooling facilities, overheating makes traditional supercom50 Computer Computer puters too unreliable for application scientists to use. Unfortunately, building exotic cooling facilities can cost as much as the supercomputer itself, and operating and maintaining the facilities costs even more. As “The Energy-Efficient Green Destiny” sidebar details, the low-power supercomputer that we developed was extremely reliable, with no unscheduled downtime in its two-year lifespan, despite residing in a dusty warehouse without cooling, humidification, or air filtration. The hourly cost of such downtime ranges from $90,000 for a catalog sales operation to nearly $6.5 million for a brokerage operation, according to Contingency Planning Research’s 2001 cost-of-downtime survey. There’s still no guarantee that the supercomputer won’t fail, as Table 1 illustrates. Total cost of ownership now exceeds initial acquisition costs. Performance at any cost The performance-at-any-cost supercomputer design paradigm is no longer feasible. Clearly, without significant change in design, the performance gains of the past two decades won’t continue. Unfortunately, performance-only metrics don’t capture improvements in power efficiency. Nonetheless, performance-only metrics derived from the Linpack benchmarks and the Standard Performance Evaluation Corp. (SPEC) code suite have significantly influenced the design of modern high-performance systems, including servers and supercomputers. Published by the IEEE Computer Society 0018-9162/07/$25.00 © 2007 IEEE Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page TMC CM-5 5 kW IBM SP ASCI White Intel Earth IBM ASCI Red 6,000 kW Simulator Blue Gene/L 850 kW 12,000 kW ~4,000 kW Fujitsu Numerical Wind Tunnel 100 kW Performance 104 Gflops 103 Gflops F High-speed electric train 10,000 kW ) R max ak ( nts k pe eme r c i a u np eq e Li er r erag pow v k a a 5 Pe Top BEMaGS Small power plant generating capacity 300,000 kW 106 Gflops 105 Gflops A Commercial data center 1,374 kW Efficiency gap tion pplica Real a Residential airconditioner 15 kW 102 Gflops 10 Gflops 1 Gflop 1993 1995 1997 1999 2001 2003 2005 2007 2009 Figure 1. Rising power requirements. Peak power consumption of the top supercomputers has steadily increased over the past 15 years. The Energy-Efficient Green Destiny As a first step toward reliable and available energyefficient supercomputing, in 2002 we built a lowpower supercomputer at Los Alamos National Laboratory. Dubbed Green Destiny, the 240-processor supercomputer took up 5 square feet (the size of a standard computer rack) and had a 3.2-kilowatt power budget (the equivalent of two hairdryers) when booted diskless.1,2 Its 101-gigaflops Linpack rating (equivalent to a 256-processor SGI Origin 2000 supercomputer or a Cray T3D MC1024-8) would have placed it at no. 393 on the TOP500 List at that time. Garnering widespread media attention, Green Destiny delivered reliable supercomputing with no unscheduled downtime in its two-year lifetime. It endured sitting in a dusty warehouse at temperatures of 85-90 degrees Fahrenheit (29-32 degrees Celsius) and an altitude of 7,400 feet (2,256 meters). Furthermore, it did so without air-conditioning, humidification control, air filtration, or ventilation. Yet despite Green Destiny’s accomplishments, not everyone was convinced of its potential. Comments ranged from Green Destiny’s being so low power that it ran just as fast when it was unplugged to the notion that no one in high-performance computing would ever care about power and cooling. However, in the past year, we’ve seen a dramatic attitude shift with respect to power and energy, particularly in light of how quickly supercomputers’ thermal power envelopes have increased in size, thus adversely impacting the systems’ power and cooling costs, reliability, and availability. The laboratory’s Biosciences Division bought a Green Destiny replica about six months after Green Destiny’s debut. In 2006, we donated Green Destiny to the division so it could run a parallel bioinformatics code called mpiBLAST. Both clusters are run in the same environment, yet half of the nodes are inoperable on the replica, which uses higher-powered processors. Hence, although the original Green Destiny was 0.150 gigahertz slower in clock speed, its productivity in answers per month was much better than the faster but often inoperable replica. Green Destiny is no longer used for computing, and resides in the Computer History Museum in Mountain View, California. References 1. W. Feng, “Making a Case for Efficient Supercomputing,” ACM Queue, Oct. 2003, pp. 54-64. 2. W. Feng, “The Importance of Being Low Power in HighPerformance Computing,” Cyberinfrastructure Technology Watch, Aug. 2005, pp. 12-21. 51 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Table 1. Reliability and availability of large-scale computing systems. System Processors ASC Q 8,192 ASC White 8,192 PSC Lemieux 3,016 Google (estimate) 450,000 Reliability and availability Mean time between interrupts: 6.5 hours, 114 unplanned outages/month Outage sources: storage, CPU, memory Mean time between failures: 5 hours (2001) and 40 hours (2003) Outage sources: storage, CPU, third-party hardware Mean time between interrupts: 9.7 hours Availability: 98.33 percent 600 reboots/day; 2-3 percent replacement/year Outage sources: Storage and memory Availability: ~100 percent Source: D.A. Reed Developing new metrics Performance-only metrics are likely to remain valuable for comparing existing systems prior to acquisition and helping drive system design. Nonetheless, we need new metrics that capture design differences in energy efficiency. For example, two hypothetical high-performance machines could both achieve 100 teraflops running Linpack and secure a high equivalent ranking on the TOP500 List. But enable smart-power-management hardware or software1,2 on one machine that can sustain performance and reduce energy consumption by 10 percent, and the TOP500 rankings remain the same. Unfortunately, metric development is fraught with technical and political challenges. On the technical side, operators must perceive the metric and its associated benchmarks as representative of the workloads typically running on the production system. On the political side, metrics and benchmarks need strong community buy-in. THE GREEN500 LIST We’ve been working to improve awareness of energyefficient supercomputer (and data center) design since the turn of the century. After interaction with government agencies, vendors, and academics, we identified a need for metrics to fairly evaluate large systems that run scientific production codes. We considered a number of methodologies for use in ranking supercomputer efficiency. To promote community buy-in, the initial Green500 List used a single metric and widely accepted workload while the intent is to extend the Green500 methodology to eventually include rankings for a suite of parallel scientific applications. The Green500 List ranks supercomputers based on the amount of power needed to complete a fixed amount of work. This effort is focused on data-center-sized deployments used primarily for scientific production codes. In contrast, the SPECPower subcommittee of SPEC is developing power-performance efficiency bench52 A BEMaGS F marks for servers running commercial production codes. The diverse types of evaluations that efforts like the Green500 and SPECPower (www.spec.org/specpower) provide will give users more choice in determining efficiency metrics for their systems and applications. Measuring efficiency In the Green500 effort, we treat both performance (speed) and power consumption as first-class design constraints for supercomputer deployments. Speed and workload. The supercomputing community already accepts the flops metric for the Linpack benchmark, which the TOP500 List uses. Although TOP500 principals acknowledge that Linpack isn’t the be-all or end-all benchmark for high-performance computing (HPC), it continues to prevail despite the emergence of other benchmarks. As other benchmark suites gain acceptance, most notably the SPEChpc3 and HPC Challenge benchmarks,4 we plan to extend our Green500 List methodology as mentioned. For now, since the HPC community seems to identify with the notion of a clearly articulated and easily understood single number that indicates a machine’s prowess, we opt to use floating-point operations per second as a speed metric for supercomputer performance and the Linpack benchmark as a scalable workload. EDn metric. There are many possibilities for performance-efficiency metrics, including circuit design’s EDn metric—with E standing for the energy a system uses while running a benchmark, D for the time to complete that same benchmark,5-8 and n a weight for the delay term. However, the EDn metrics are biased when applied to supercomputers, particularly as n increases. For example, with large values for n, the delay term dominates so that very small changes in execution time impact the metric dramatically and render changes in E undetectable in comparisons. Flops per watt. For the Green500 List, we opted to use flops per watt for power efficiency. However, this metric might be biased toward smaller supercomputing systems. A supercomputer’s wattage will scale (at least) linearly with the number of compute nodes while the flops performance will scale (at most) linearly for embarrassingly parallel problems and sublinearly for all other problems. This implies smaller systems would have better ratings on such a scale. Nonetheless, flops per watt is easy to measure and has traction in the scientific community. Furthermore, we can reduce the bias toward small systems by ranking systems that first achieve a minimum performance rating. We simply set a minimum flops threshold for entry into the Green500 List and allow bigger super- Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page computers to rerun their Linpack benchmark, if desired, to meet this minimum threshold and obtain the corresponding power consumption during a benchmark’s rerun. That is, the Green500 List ranks supercomputers based on the amount of power needed to complete a fixed amount of work at a rate greater than or equal to the minimum flops threshold. Measuring power consumption Even after choosing the benchmark and power-efficiency metric, issues surrounding the selection of the flopsper-watt metric for a given supercomputer remained unresolved. First, starting with the metric’s numerator, what will be the minimum flops threshold for entry into the Green500 List? We intend to use the flops rating that the no. 500 supercomputer achieves on the latest TOP500 List as the Green500 List’s minimum flops threshold. With the numerator addressed, this leaves wattage as the denominator. Surprisingly to some, the denominator for flops per watt might be more difficult to determine, since there are many permutations for what we could measure and report. For example, we could • measure the entire supercomputer’s power consumption, • measure a single supercomputer node and extrapolate it to the entire supercomputer, or • use the manufacturers’ advertised peak power numbers (as we used in Figure 1). Measuring wattage for a supercomputer the size of a basketball court is difficult. However, using advertised peak power numbers could result in overinflation of the power numbers. We suggest measuring a single compute node’s power consumption and multiplying by the number of compute nodes (loosely defining a node as an encased chassis, whether the chassis has the form factor of a standard 1U server or an entire rack). Power meters. To measure power consumption, we propose using a power meter that can sample consumption at granularities of one second or less. The digital meters range in capability from the commodity Watts Up? Pro (www.wattsupmeters.com) to the industrialstrength Yokogawa WT210/WT230 (http://yokogawa. com/tm/wtpz/wt210/tm-wt210_01.htm). Figure 2 shows a high-level diagram of how a digital power meter measures a given system under test (single compute node) via a common power strip and logs the measurements to a profiling computer. Duration. We also need to address how long to measure power and what we should record and report. Given the meters’ recording capabilities, we suggest measuring and recording power consumption for the duration of the Linpack run and using the average power consumption over the entire run. Coupling average power consumption with the Linpack run’s execution Profiling computer System under test Digital power meter Power strip A BEMaGS F Wall power outlet Figure 2. Power-measurement infrastructure. A digital power meter measures a system under test via a common power strip and logs the measurements to a profiling computer. time adds inferred overall energy consumption, where energy is average power multiplied by time. Cooling. Finally, we considered whether to include cooling-facility power consumption in the measurement. We decided against inclusion because the Green500 List is intended to measure the supercomputer’s power efficiency, rather than cooling systems (which vary widely in power efficiency). Even if we considered cooling for the supercomputer under test, it would be difficult to break out and measure cooling’s specific contribution for one supercomputer, given that cooling facilities are designed to support all the machines in a given machine room. TOP500 VERSUS GREEN500 Table 2 presents the Green500 and TOP500 rankings of eight supercomputers, as well as their flops ratings and their peak power usage. This list also shows the results of using the flops-per-watt metric for these supercomputers with their peak performance number (for peak power efficiency) and their Linpack performance number (for actual power efficiency). As mentioned, using peak power numbers for comparisons isn’t optimal. Nonetheless, the relative comparisons using peak power numbers are useful to gauge power-efficiency progress. Beginning with the November 2007 Green500 List, we’ll use metered measurements in rankings whenever available. As the list matures, we anticipate metering and verifying all measurements. Various presentations, Web sites, and magazine and newspaper articles provide the source for these peak power numbers. For the IBM Blue Gene/L supercomputer at Lawrence Livermore National Laboratory (LLNL), the TOP500 wiki reports 1.5 MW as its peak power consumption. LLNL’s Web site reports that 7.5 MW is needed to power and cool ASC Purple, while Eurekalert estimates it uses 8 MW. According to LLNL, for every watt of power the system consumes, 0.7 watts of power is required to cool it. Hence, the power required to merely run ASC Purple would be between 4.4 and 4.7 MW, which matches the 4.5 MW number provided in a presentation at a Blue Gene/L workshop. Jaguar at Oak Ridge National Laboratory is a hybrid system consisting of 56 XT3 cabinets and 68 53 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Table 2. June 2007 Green500 and TOP500 rankings. Green500 rank (power efficiency) Supercomputer 1 2 3 4 5 6 7 8 Blue Gene/L MareNostrum Jaguar System X Columbia ASC Purple ASC Q Earth Simulator Peak performance (Gflops) Linpack performance (Gflops) 367,000 94,208 119,350 20,240 60,960 92,781 20,480 40,960 280,600 62,630 101,700 12,250 51,870 75,760 13,880 35,860 Peak power (kW) Peak power efficiency (Mflops/W) Green500 rank (peak power efficiency Power efficiency (Mflops/W) TOP500 rank 1,200 1,344 2,087 310 2,000 4,500 2,000 7,680 305.83 70.10 57.19 65.29 30.48 20.62 10.24 5.33 1 2 4 3 5 6 7 8 233.83 46.60 48.73 39.52 25.94 16.84 6.94 4.67 1 9 2 71 13 6 62 20 Table adapted from a figure provided by NXP Semiconductors. XT4 cabinets. The peak power consumption of an XT3 cabinet is 14.5 kW while the XT4 cabinet is 18.75 kW, as per Cray datasheets. Thus, the aggregate peak power of Jaguar is about 2 MW. The 4,800-processor MareNostrum debuted fifth on the June 2005 TOP500 List with an estimated 630-kW power budget to run the machine. More recently, Barcelona Supercomputing Center’s MareNostrum was upgraded and expanded into a 10,240-processor BladeCenter JS21 cluster. If we extrapolate from the original MareNostrum’s 630-kW power budget, the 10,240processor MareNostrum would have a power budget of 1.3 MW. For the Columbia supercomputer at NASA Ames Research Center, the reported power usage just to run the system is 2 MW. The thermal design power of Itanium-2 processors is 130 watts, so it takes 1.33 MW to run the 10,240 processors in the Columbia. Therefore, 2 MW seems reasonable if Columbia’s other components use only 700 kW of power, consistent with our Itanium-based server’s power profile. Powering and cooling Japan’s 5,120-processor Earth Simulator requires 11.9 MW, enough to power a city of 40,000 and a 27,000-student university. The Earth Simulator configures the 5,120 processors into 640 eight-way nodes, where each eight-way node uses 20 kilovolt-amperes. Assuming a typical power-factor conversion of 0.6, each node then consumes 20 kVA 0.6 = 12 kW. Thus, power consumption for the entire 640node Simulator is 640 12 kW = 7,680 kW, leaving 4,220 kW for cooling. The power budgets for ASC Q and ASC White run at approximately 2 MW, while System X at Virginia Tech consumes a paltry 310 kW, as measured directly from System X’s power distribution units. As Table 2 shows, despite its large size, Blue Gene/L is the only custom lowpower supercomputer among the TOP500. It’s routinely the highest-ranking supercomputer on both the TOP500 54 and Green500 lists, with a performance-power ratio that’s up to two orders of magnitude better than the other supercomputers in Table 2. The power efficiencies of MareNostrum (semicommodity) and System X (commodity) are 2.5 times better than the other supercomputers, and this ranked them second and fourth on the June 2007 Green500 List, as shown in Table 2. Interestingly, Apple, IBM, and Motorola’s commodity PowerPC processor drives both of these power-efficient supercomputers. On the other hand, ASC Purple, which ranked sixth on that TOP500 list, is also based on the PowerPC processor, albeit the Power5, its higher-powered relative. Power5 ultimately contributes to ASC Purple’s lower power efficiency and its sixth-place ranking on the 2007 Green500. OPERATIONAL COSTS AND RELIABILITY Power consumption has become an increasingly important issue in HPC. Ignoring power consumption as a design constraint results in an HPC system with high operational costs and diminished reliability, which often translates into lost productivity. With respect to high operational costs, ASC Purple has a 7.5-MW appetite (approximately 4.5 MW to power the system and 3 MW for cooling). With a utility rate of 12 cents per kW/hour, the annual electric bill for this system would run nearly $8 million. If we scaled this architecture to a petaflops machine, powering up and cooling down the machine would require approximately 75 MW. The system’s annual power bill could run to $80 million, assuming energy costs remained the same. Table 1 shows that the reliability and availability of large-scale systems, ranging from supercomputers to a large-scale server farm, is often measured in hours. Further scaling of such supercomputers and data centers would result in several failures per minute.9 This diminished reliability results in millions of dollars per hour in lost productivity. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F In light of the above, the HPC community could use an EnergyGuide sticker, such as the Green Destiny sticker shown in Figure 3. The community could also use studies showing that annual server power and cooling costs are approaching annual spending on new machines. The HPC community needs a Green500 List to rank supercomputers on speed and power requirements and to supplement the TOP500 List. Vendors and system architects worldwide take substantial pride and invest tremendous effort toward making the biannual TOP500 List. We anticipate that the Green500 List effort will do the same and encourage the HPC community and operators of Internet data centers to design more power-efficient supercomputers and large-scale data centers. For the latest Green500 List, visit www.green500.org. ■ Acknowledgments We thank David Bailey, Jack Dongarra, John Shalf, and Horst Simon for their support. Intel, IBM, Virginia Tech, the US Department of Energy, and the NSF (CNS 0720750 and 0615276; CCF 0709025 and 0634165) sponsored portions of this work. We also acknowledge colleagues who suggested a Green500 List after an April 2005 keynote at the IEEE International Parallel & Distributed Processing Symposium and a follow-up talk at North Carolina State University. This article is dedicated to those who lost their lives in the 16 April 2007 tragedy at Virginia Tech. __________ Figure 3. EnergyGuide sticker for Green Destiny. Such a sticker could remind those in the HPC community of a computer’s energy use and hourly operating costs. References 1. C. Hsu and W. Feng, “A Power-Aware Run-Time System for High-Performance Computing,” Proc. ACM/IEEE SC Conf. (SC05), IEEE CS Press, 2005, p. 1. 2. K.W. Cameron et al., “High-Performance, Power-Aware Distributed Computing for Scientific Applications,” Computer, Nov. 2005, pp. 40-47. 3. M. Mueller, “Overview of SPEC HPC Benchmarks,” BOF presentation, ACM/IEEE SC Conf. (SC06), 2006. 4. J. Dongarra and P. Luszczek, Introduction to the HPC Challenge Benchmark Suite, tech. report, Univ. Tennessee, 2004; www.cs.utk.edu/~luszczek/pubs/hpcc-challenge-intro.pdf. 5. A. Martin, “Towards an Energy Complexity of Computation,” Information Processing Letters, vol. 77, no. 2-4, 2001, pp. 181-187. 6. D. Brooks and M. Martonosi, “Dynamically Exploiting Narrow Width Operands to Improve Processor Power and Performance,” Proc. 5th Int’l Symp. High-Performance Computer Architecture, IEEE CS Press, 1999, p. 13. 7. R. Gonzalez and M. Horowitz, “Energy Dissipation in General-Purpose Microprocessors,” IEEE J. Solid-State Circuits, Sept. 1996, pp. 1277-1284. 8. A. Martin, M. Nystrm, and P. Penzes, ET2: A Metric for Time and Energy Efficiency of Computation, Kluwer Academic Publishers, 2002. 9. S. Graham, M. Snir, and C. Patterson, eds., Getting Up to Speed: The Future of Supercomputing, Nat’l Academies Press, 2005. Wu-chun Feng is an associate professor of computer science and electrical and computer engineering at Virginia Tech. His research interests are high-performance networking and computing. Feng received a PhD in computer science from the University of Illinois at Urbana-Champaign. He is a senior member of the IEEE Computer Society. Contact him at __________ feng@cs.vt.edu. Kirk W. Cameron is an associate professor of computer science at Virginia Tech and director of its Scalable Performance Lab. His research interests are power and performance in high-performance applications and systems. Cameron received a PhD in computer science from Louisiana State University. He is a member of the IEEE Computer Society. Contact him at _____________ cameron@cs.vt.edu. 55 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page C O V E R A BEMaGS F F E A T U R E Life Cycle Aware Computing: Reusing Silicon Technology John Y. Oliver, Cal Poly San Luis Obispo Rajeevan Amirtharajah and Venkatesh Akella, University of California, Davis Roland Geyer and Frederic T. Chong, University of California, Santa Barbara Despite the high costs associated with processor manufacturing, the typical chip is used for only a fraction of its expected lifetime. Reusing processors would create a “food chain”of electronic devices that amortizes the energy required to build chips over several computing generations. T he past decade has seen unprecedented growth in the number of electronic devices available to consumers. Many of these devices, from computers to set-top boxes to cell phones, require sophisticated semiconductors such as CPUs and memory chips. The economic and environmental costs of producing these processors for new and continually upgraded devices are enormous. Because the semiconductor manufacturing process uses highly purified silicon, the energy required is quite high—about 41 megajoules (MJ) for a dynamic random access memory (DRAM) die with a die size of 1.2 cm2.1 To illustrate the macroeconomic impact of this energy cost, Japan’s semiconductor industry is expected to consume 1.7 percent of the country’s electricity budget by 2015.2 Approximately 600 kilograms of fossil fuels are needed to generate enough energy to create a 1-kilogram semiconductor.3 Furthermore, according to chip consortium Sematech, foundry energy consumption also continues to increase.4 In terms of environmental impact, 72 grams of toxic chemicals are used to create a 1.2 cm2 DRAM die. The semiconductor industry manufactured 28.4 million cm2 of such dies in 2000, which translates to 1.7 billion kilograms of hazardous material.2 Due to the increasing number of semiconductor devices manufactured each year, semiconductor disposal costs are likewise increasing. Despite these costs, the typical processor is used for only a fraction of its expected lifetime. While rapid tech- 56 Computer Computer nological advances are quickly making silicon obsolete, chips could be removed from recycled electronics and reused for less demanding computing tasks. A processor reuse strategy would create a “food chain” of computing devices that amortizes the energy required to build processors—particularly low-power, embedded processors—over several computing generations. PROCESSOR LIFETIME ENERGY CONSUMPTION The lifetime energy consumption of a processor or memory chip can be expressed as the sum of the • manufacturing energy cost, including the creation of silicon wafers, the chemical and lithography processes, and chip assembly and packaging; and • utilization energy cost. A comparative analysis of these two components reveals that the energy required to manufacture a processor can dominate the energy consumed over the processor’s lifetime. Manufacturing energy cost Semiconductor manufacturing involves many steps, from crystal growth to dicing to packaging. Total energy cost can be expressed as Emanufacturing = Edie + Eassembly. Edie is the energy required to manufacture the die of the processor or memory chip and includes wafer growth, epitaxial layering, applying photo resists, etching, Published by the IEEE Computer Society 0018-9162/07/$25.00 © 2007 IEEE Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer 100 100 80 80 (a) F 50 mm2 in 2006 40 (b) 2012 2009 2006 2012 2009 2006 2012 2009 2012 (32 nm) 2006 0 2009 (45 nm) BEMaGS 20 0 2006 (65 nm) 100 mm2 in 2006 A Packaging and assembly cost Die Cost 2012 20 150 mm2 in 2006 60 2009 200 mm2 150 mm2 100 mm2 50 mm2 40 200 mm2 in 2006 2006 60 Energy (MJ) Yield (percent) Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Year Figure 1. Semiconductor yield and manufacturing energy costs over time. (a) Shrinking the processor increases yield, which (b) decreases manufacturing energy costs over subsequent generations. implantation/diffusion, and managing these procedures. Eassembly represents the cost to assemble the chip and includes wafer testing, dicing, bonding, encapsulation, and burn-in testing. Based on this simple formula, the authors of a recent study1 made several assumptions about the manufacturing energy cost for any CMOS-based semiconductor. First, they assumed that the energy required to manufacture a 1.2 cm2 processor at any lithographical level is the same—thus, the energy costs of manufacturing a 1.2 cm2 DRAM die and 1.2 cm2 processor die are identical. Another assumption is that the manufacturing energy required is proportional to the semiconductor die area (Edie = 1/yield area), so that a 0.6 cm2 processor requires half as much energy for die manufacture as a 1.2 cm2 die, adjusted for yield. Finally, they assumed that the assembly energy cost is a constant 5.9 MJ, regardless of the die size. For a 1.2 cm2 DRAM chip, Emanufacturing = 41 MJ, Edie = 35.1 MJ, and Eassembly = 5.9 MJ. As part of their manufacturing energy analysis, the researchers employed the SUSPENS (Stanford University System Performance Simulator) yield model.5 According to this model, yield = eD0 area, with the D0 constant taken from the International Technology Roadmap for Semiconductors.6 Figure 1a shows the yield curves for four hypothetical processors over time. Shrinking the processor clearly increases yield. For example, the yield for a 200 mm2 processor in 2006 is 60 percent; the same processor, shrunk using 2012 technology, has a yield over 80 percent. The manufacturing cost for subsequent generations of processors thus has the potential to decrease due to shrinking transistor geometry. Figure 1b demonstrates the energy required to manufacture a processor with fixed functionality over time. The energy savings in subsequent years is due to shrinking transistor geometries and yield improvements. Processors with larger dies have a higher percentage of energy savings because packaging costs are a smaller portion of the overall manufacturing cost. Also, in the extreme case, shrinking a processor might make it pad limited. The physical dimensions of a pad are unlikely to shrink far below 60 m on a side.7 We believe that the projected figures shown in Figure 1b are on the conservative side. The data the researchers used is from a 4-inch wafer fab, and modern 12-inch wafers require more energy per unit area to process.4 In addition, many modern semiconductor processes have more layers than the process used in the study. The amount of energy required to manufacture a processor die is clearly considerable. A 300-mm wafer uses 2 gigajoules of energy, which is roughly the amount contained within 200 gallons of gasoline. The good news is that the total manufacturing energy cost diminishes with every process shrink. Unfortunately, packaging and assembly costs are relatively fixed. Utilization energy cost A processor’s utilization energy cost can be determined by simply multiplying its power consumption by the time it is operational. For example, the Intel XScale PX273 consumes 0.77 watts of power in full operation.8 Assuming that an XScale-based PDA is used two hours per day 365 days per year, the PX273 consumes just over 2 MJ of energy annually. One factor that can impact a processor’s power consumption is the manufacturing process technology. A benefit of shrinking transistor geometry is that circuits’ switching capacitance decreases with each shrink. For low-end cell phones and other devices with relatively fixed performance, processor power consumption may benefit from process shrinks unless leakage current becomes problematic. Higher amounts of leakage make processor reuse a more attractive solution than upgrading to a new process technology, as the processors manufactured with older process technologies will have lower leakage current. 57 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F PROCESSOR FOOD CHAIN 140 Energy (MJ) 120 100 80 60 No upgrade Upgrade every 2 years Upgrade every 4 years 40 20 0 Energy (MJ) (a) 900 800 700 600 500 400 300 200 100 0 1 2 3 4 5 6 Time (years) 7 8 9 No upgrade Upgrade every 2 years Upgrade every 4 years 10 Mobile device processors are typically used for only a fraction of their designed lifetime. “Computer chips can operate for 80,000 hours, and usually machines are thrown out after 20,000 hours,” observed Guardian columnist John Keeble. “However, at the moment, 60 percent of chips cannot be reused because of their specialized functions.”9 To facilitate reuse, researchers could standardize embedded processor footprints for a wide range of embedded devices. In addition, instead of reusing a processor in the same device, it could serve a next-generation device with lower performance requirements. Researchers also could apply power-savings techniques like voltage scaling, given the secondary device’s lower computational demand and corresponding operational frequency and voltage. Example: ARM9 To illustrate how a food chain of electronic devices could reuse a processor, consider the ARM9 processor, which is featured in the Alpine 1 2 3 4 5 6 7 8 9 10 Blackbird PMD-B100 and Sell GPS-350A auto(b) Time (years) motive navigation systems. The ARM9 implementation in these systems runs at 266 MHz. Once Figure 2. Potential benefits of processor reuse. (a) Upgrading a 1-W the navigation system is recycled, the processor can processor does not improve the lifetime energy consumption for at be removed and placed into a mobile phone like least 10 years, making processor reuse an attractive alternative. (b) the Sony Ericsson P800, which uses a similar For processors that use more power—in this case, 20 W—upgrading ARM9 processor running at 156 MHz. When this with newer, more efficient technology makes sense. phone is recycled, the processor can in turn be put into a Nintendo DS portable game system, which PROCESSOR REUSE uses an ARM9 running at 77 MHz. Figure 2a illustrates how processor reuse minimizes Table 1 compares the lifetime energy consumption of the lifetime energy consumption of a processor that uses a processor reuse strategy with a strategy that uses new 1 W of power. The two- and four-year upgrade curves processors in this chain of devices. These results assume increase every two and four years, depicting the high that the automotive navigation system is used one hour energy cost of manufacturing the processors. These per day, the mobile phone three hours per day, and the results are based on the assumption that the processor Nintendo DS game system two hours per day, every day has a die area of 1.2 cm2, is operated three hours every for three years, before being recycled. day, and is dormant (but still leaking) when not in use. Note that manufacturing energy constitutes a large Processors with a 1-W rating or less clearly should not portion of the processors’ lifetime energy consumption. be upgraded with new processors to reduce their life- In addition, the manufacturing energy cost of chips in time energy consumption. On the other hand, as Figure 2009 and 2012 for the new-processor chain decreases 2b shows, upgrading is a viable option to minimize life- only slightly. Some decrease is expected, as the die size time energy consumption for a higher-power proces- shrinks in each generation, but the decrease is limited by the fixed amount of energy required to assemble the sor—in this case, a 20-W processor. To minimize lifetime energy consumption, it makes processors and the fact that pad size is unlikely to scale sense to reuse a processor when it uses 100 kJ of energy with technology.7 Also noteworthy is that reused procesper day or less. Assuming that upgrading occurs in three- sors have a higher utilization cost than new ones. The year cycles and the device containing the processor is increase is small, but it could be important for severely used three hours per day, this is roughly equivalent to power-constrained devices. the energy a 10-W processor consumes. For perspective, This study neglects the energy required to reclaim a 100 kJ of energy is a bit less energy than is contained processor, but processor reuse has other benefits that within a fully charged laptop battery, or about the same counterbalance this including reduced disposal costs and amount in 10 cell-phone batteries. decreased toxic chemical use. Also, a processor recla58 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer A BEMaGS Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page F mation infrastructure already exists, albeit in a black market fashion.10 Table 1. Lifetime energy consumption: processor reuse versus using new processors. New processor every 3 years Processor reused every 3 years Performance requirements Manufacturing Manufacturing Utilization Utilization Despite its potential benefits, processor reuse poses both technical and economic obstacles. Home tool White good MP3 player Toy Home stereo Portable game Printer Cell phone Set-top box PDA Automotive navigation system REUSABLE PROCESSOR CHALLENGES Digital video camera BDTImark Figure 3 shows the BDTImark energy cost (MJ) energy cost (kJ) Year energy cost (MJ) energy cost (kJ) performance of a variety of electronic devices. The blue bars 6.88 36.92 2006 6.88 36.92 indicate devices that commonly 0 36.92 2007 0 36.92 use specialized hardware to 0 36.92 2008 0 36.92 accelerate processing and thus 6.40 28.87 2009 0 153.74 may have considerably higher 0 28.87 2010 0 153.74 requirements than indicated. 0 28.87 2011 0 153.74 The opportunities for processor 6.29 4.55 2012 0 50.59 reuse are evident: A processor 0 4.55 2013 0 50.59 used in a particular device 0 4.55 2014 0 50.59 should be capable of handling 19.57 211.02 Total 6.88 723.75 the processing required by all 19.78 MJ Lifetime 7.60 MJ devices to the right of it in Figure 3. For example, the processor from a PDA could be reused in an automo- Technical challenges bile navigation system. In order to facilitate processor reuse, it will be necesOver time, the range of performance requirements sary to support some circuit flexibility on the die of a should continue to grow as the functionality of these reusable processor. To ascertain how much circuit area devices expands. However, given the ever-present need overhead a reusable processor can tolerate, we comfor low-end processing, a food chain of applications will pared the manufacturing and utilization energy costs for always exist in some form. a strategy that uses new processors every three years with one that uses a single processor every three years Battery-constrained devices for a total lifetime of nine years. Subtracting the energy Because reused processors are manufactured with for the latter strategy from that for the former, we then process technology that is potentially several years converted this energy differential to an amount of allowolder than state of the art, reused processors have able “additional area” on a reusable processor—that is, higher utilization energy requirements than new ones. we assumed this extra circuitry consumes the same Voltage scaling can mitigate this disadvantage. A reused processor that is 2,500 higher up on the food chain will have a higher peak performance than what is 2,000 required by a device that is lower on the food chain. Scaling back the frequency, 1,500 and therefore the voltage of the reused processor, significantly reduces its energy requirements. 1,000 In addition, many mobile devices already have adequate battery life. For example, 500 the Nintendo DS game system can run up to 10 hours on a single charge. If the system 0 is used two hours per day, it would have to be recharged once every five days with a new processor but potentially once every four days with a reused processor. Figure 3. BDTImark performance of various electronic devices. A processor used in a particular device should be capable of handling the processing required by all devices to the right of it. 59 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page ARM920T ARM920T ARM920T XScale XScale XScale XScale ARM9 ARM7 Additional area (mm2) 40 35 30 25 20 15 10 2 4 6 8 10 Utilization (hours/day) BEMaGS F Free-market economic incentives, however, might be insufficient. Environmental protection is often within the purview of public policy. European Union directives to reduce hazardous waste in electronic devices, such as the Restriction of Hazardous Substances (RoHS 2002/95/ EC),11 have effectively led all major chipmakers to adopt plans such as moving to lead-free solder. More relevant to processor reuse, the Kyoto Protocol to the United Nations Framework Convention on Climate Change establishes a market economy for greenhouse gas emissions that creates an added financial incentive to reduce energy usage and create carbon-neutral products.12 12 Figure 4. Chip area available for additional circuitry on three processor reuse chains while still maintaining lifetime energy efficiency. amount of active energy per mm2 as the processor core. Figure 4 shows the additional circuit area that can support processor reuse while reducing the processor’s lifetime energy consumption. This allowable area budget clearly depends on the processor’s utilization. Processors used less frequently utilize less power and therefore have a higher allowable area budget. The top line in Figure 4 represents a chain of three processors with capabilities similar to those of an ARM920T. The higher the processor’s utilization, the less processor area that can be used for reuse support. For higher-power chips, such as the Intel XScale series illustrated by the bottom line, the reuse-support area decreases significantly. For reuse chains that involve devices with subsequently smaller computational requirements, the area available for reuse support is quite high due to low utilization energy. This is shown by the middle line, which is a reuse strategy based on an XScale in the first generation, ARM9 in the second generation, and ARM7 processor in the third generation. Overall, the additional area for supporting reuse is quite large: An XScale processor core is about 20 mm2 in 130nm technology. Economic challenges A major obstacle to processor reuse is that chipmakers would not profit from this strategy unless they become actively involved in salvaging and reselling operations. On the other hand, they would suffer financially only if third parties sold reused chips that competed with the manufacturer’s new offerings. Conceptually, the easiest solution would be for chipmakers to charge a premium price for reusable processors that owners of the product containing the chip could recover when returning the product for recycling. Another option would be to credit the chipmaker when one of its processors is reused. 60 A M oore’s law has led to a disposable-chip economy with increasingly severe economic and environmental costs. The energy required to manufacture low-power, embedded processors is so high that reusing them can save orders of magnitude of lifetime energy per chip. Processor reuse will require innovative techniques in reconfigurable computing and hardwaresoftware codesign as well as governmental policies that encourage silicon reuse, but the potential benefits to society will be well worth the effort. ■ References 1. E.D. Williams, R.U. Ayres, and M. Heller, “The 1.7 Kilogram Microchip: Energy and Material Use in the Production of Semiconductor Devices,” Environmental Science and Technology, vol. 36, no. 24, 2002, pp. 5504-5510. 2. R. Kuehr and E. Williams, eds., Computers and the Environment: Understanding and Managing Their Impacts, Kluwer Academic Publishers, 2003. 3. E.D. Williams, “Environmental Impacts of Microchip Manufacture,” Thin Solid Films, vol. 461, no. 1, 2004, pp. 2-6. 4. “ISMI Study Finds Significant Cost Savings Potential in Fab Energy Reduction,” Sematech news release, 22 Dec. 2005; www.sematech.org/corporate/news/releases/20051222a.htm. 5. H.B. Bakoglu, Circuits, Interconnections, and Packaging for VLSI, Addison-Wesley, 1990. 6. International Technology Roadmap for Semiconductors: 2005 Edition—System Drivers, ITRS, 2005; www.itrs.net/Links/ 2005ITRS/SysDrivers2005.pdf. ___________________ 7. J. Courtney, N. Aldahhan, and M. Engloff, “The ProbeCentric Future of Test,” presentation, 2006 Southwest Test Workshop; www.swtest.org/swtw_library/2006proc/PDF/ S06_01_IMSI-SEMATECH.pdf. ____________________ 8. L.T. Clark et al., “Standby Power Management for a 0.18 m Microprocessor,” Proc. 2002 Int’l Symp. Low Power Electronics and Design, ACM Press, 2002, pp. 7-12. 9. J. Keeble, “From Hackers to Knackers,” The Guardian, supplement online, 21 May 1998; http://online.guardian.co.uk. 10. M. Pecht and S. Tiku, “Bogus: Electronic Manufacturing and Consumers Confront a Rising Tide of Counterfeit Electronics,” IEEE Spectrum, vol. 43, no. 5, 2006, pp. 37-46. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page 11. “Directive 2002/95/EC of the European Parliament and of the Council of 27 January 2003 on the Restriction of the Use of Certain Hazardous Substances in Electrical and Electronic Equipment,” Official J. European Union, vol. 37, 2003, pp. 19-23; www.interwritelearning.com/rohs_compliance.pdf. 12. “Kyoto Protocol to the United Nations Framework Convention on Climate Change,” United Nations, 1998; http://unfccc. int/resource/docs/convkp/kpeng.pdf. ______________________ John Y. Oliver is an assistant professor in the Department of Computer Engineering at California Polytechnic State University, San Luis Obispo. His research interests include computer architecture, reliability in computing, and sustainable computing. Oliver received a PhD in computer engineering from the University of California, Davis. He is a member of the IEEE and the ACM. Contact him at jyoliver@calpoly.edu. ______________ Rajeevan Amirtharajah is an assistant professor in the Department of Electrical and Computer Engineering at the University of California, Davis. His research interests include digital and mixed-signal circuit design, low-power signal processing, and system architectures. Amirtharajah received a PhD in electrical engineering and computer science from the Massachusetts Institute of Technology. He is a member of the IEEE and the American Association for the Advancement of Science. Contact him at _______ amirtharajah@ece.ucdavis.edu. ______________ Venkatesh Akella is a professor in the Department of Electrical and Computer Engineering at the University of California, Davis. His current research interests include FPGAs, computer architectures, and embedded systems with an emphasis on low power and reconfigurability. Akella received a PhD in computer science from the University of Utah. He is a member of the ACM. Contact him at ______ akella@ ucdavis.edu. ________ Roland Geyer is an assistant professor in the Donald Bren School of Environmental Science and Management at the University of California, Santa Barbara. His research focuses on the life cycle of manufactured goods, the environmental and economic potential of reuse and recycling activities, and the evolution of green business plans. Geyer received a PhD in engineering from the University of Surrey, Guildford, UK. Contact him at geyer@bren.ucsb.edu. ______________ Frederic T. Chong is a professor in the Department of Computer Science as well as director of the Computer Engineering Program at the University of California, Santa Barbara. His research interests include next-generation embedded architectures, quantum computing architectures, and hardware support for system security. Chong received a PhD in electrical engineering and computer science from the Massachusetts Institute of Technology. He is a member of the ACM. Contact him at _____________ chong@cs.ucsb.edu. A BEMaGS F How to Reach Computer Writers We welcome submissions. For detailed information, visit www.computer.org/computer/author.htm. News Ideas Contact Lee Garber at ________________ lgarber@computer.org with ideas for news features or news briefs. Products and Books Send product announcements to ____________ developertools@ computer.org. Contact computer-ma@computer.org ____________________ with book announcements. Letters to the Editor Please provide an e-mail address with your letter. Send letters to computer@computer.org. __________________ On the Web Explore www.computer.org/computer/ for free articles and general information about Computer magazine. Magazine Change of Address Send change-of-address requests for magazine subscriptions to address.change@ieee.org. __________________ Make sure to specify Computer. Missing or Damaged Copies If you are missing an issue or received a damaged copy, contact ____________________ membership@computer.org. Reprint Permission To obtain permission to reprint an article, contact William Hagen, IEEE Copyrights and Trademarks Manager, at whagen@ieee.org. ____________ To buy a reprint, send a query to __________________ computer@computer.org. 61 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F REPORT TO MEMBERS Land Voted 2008 Computer Society President-Elect New vice presidents and Board of Governors members also chosen I EEE Computer Society members Publications Board, the Educational recently selected Susan (Kathy) Activities Board, the Conferences and Land, CSDP, of Northrop GrumTutorials Board, the Standards Activiman Information Technology, to ties Board, the Technical Activities serve as the Society’s presidentBoard, the Chapter Activities Board, elect for 2008. and the Student Activities Board. Land is currently the IEEE Computer All appointed Society vice presidents Society’s second vice president for stanalso serve as nonvoting members of dards activities. She is chair of the IEEE the Board of Governors. Holding votComputer Society Software Engineering positions on the Board are the 2008 IEEE Computer Society ing Portfolio oversight committee, a president, past president, presidentPresident-Elect Susan (Kathy) Land, member of the IEEE Computer Society elect, and the first and second vice CSDP, will introduce efforts to ensure International Design Competition presidents. Additional nonvoting that IEEE Computer Society products Committee, and chair of the Computer members are the Society’s staff execuremain relevant to the marketplace. Society Technical Achievement Award tive director, the editor in chief of subcommittee. Land is the author of Computer, and the IEEE directors for Jumpstart CMM/CMMI Software divisions V and VIII—the Computer Process Improvement: Using IEEE Society’s elected representatives on the Software Engineering Standards (John IEEE Board of Governors. Wiley & Sons, 2005). She is coauthor of Practical Support for CMMI-SW BOARD OF GOVERNORS ADDS Software Project Documentation: SEVEN NEW MEMBERS Using IEEE Software Engineering In the 2007 Society election, which Standards (John Wiley & Sons, 2005) closed in early October, voters also and Practical Support for ISO 9001 cast ballots to fill seven openings on 2008 President Rangachar Kasturi is Software Project Documentation: the IEEE Computer Society Board of working to build a stronger and more Using IEEE Software Engineering Governors. The full Board consists of agile organization. Standards (John Wiley & Sons, 2006). 21 members. Each year, seven new or Candidates elected to the Computer returning members are elected to Society presidency serve a three-year term in a leadership serve three-year terms. Members chosen for 2008-2010 role. After serving a year as president-elect under 2008 terms are André Ivanov, Phillip Laplante, Itaru Mimura, president Rangachar Kasturi, Land will assume the duties Jon Rokne, Christina Schober, Ann Sobel, and Jeffrey of Society president in 2009. Following her term as pres- Voas. Many of the successful candidates have had recent ident, Land will continue to be an active Society leader Board of Governors experience. Elected officers volunteer their time and talents to furin 2010 as past president. ther the Society’s goals and to elevate the profile of the NEW VICE PRESIDENTS ELECTED computing profession in general. Society officers take a George Cybenko was elected 2008 first vice president, lead role in promoting new publications, educational while Michel Israel topped the balloting for 2008 second efforts, technical focus groups, and international standards vice president. Each will serve as chair of one of the sev- that help Computer Society members attain career goals. eral Computer Society boards. The sitting president also The Computer Society mailed 73,571 ballots to memappoints vice presidents to complement the two elected bers in the 2007 election. Of the 7,106 ballots cast—a VPs as leaders of individual Society activities boards: the return rate of 9.66 percent—4,264 were submitted via 62 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Table 1.These new officers will begin serving the IEEE Computer Society on 1 January 2008. Office Officer 2008 president-elect/2009 president 2008 first vice president 2008 second vice president 2008-2010 terms on the Board of Governors Susan (Kathy) Land, CSDP George Cybenko Michel Israel André Ivanov Phillip Laplante Itaru Mimura Jon Rokne Christina Schober Ann Sobel Jeffrey Voas Number of votes Percent 3,808 3,721 3,723 3,379 3,511 3,116 3,243 4,275 5,143 3,605 53.59* 52.36* 52.39* 6.79 7.06 6.26 6.52 8.59 10.34 7.25 Table 2. The full ballot for the 2007 election also included the following candidates. Office Officer 2007 president-elect/2008 president 2007 first vice president 2007 second vice president 2008-2010 terms on the Board of Governors James Isaak Sorel Reisman Antonio Doria Alfredo Benso Fernando Bouche Joseph Bumblis Hai Jin Gerard Medioni Raghavan Muralidharan Number of votes Percent 3,163 3,176 3,089 3,023 2,533 2,924 2,913 2,825 2,854 44.51* 44.69* 43.47* 6.08 5.09 5.88 5.86 5.68 5.06 *Percentage reflects only ballots cast for this office. A BEMaGS F the Web, 2,824 were mailed in, and 18 were cast by fax. Table 1 shows the breakdown of votes cast for each officer. The full ballot for the 2007 election also included the candidates listed in Table 2. LEADERS SERVE MEMBERS Each year, Society members vote for the next year’s president-elect, first and second vice presidents, and seven members of the IEEE Computer Society Board of Governors. The Society president and vice presidents each serve a one-year active term, while the 21 Board of Governors members serve three-year terms, rotating in three groups of seven. The three presidents—incoming, active, and outgoing—work together in setting policy and making operational decisions. The active Society president is responsible for heading the annual Board of Governors meetings and for addressing major issues that affect the Computer Society during the year. NOMINATE A CANDIDATE Any Computer Society member can nominate candidates for Society offices. Most members are also eligible to run for a seat on the Board of Governors. Candidates for other offices must be full members of the IEEE and must have been members of the Computer Society for at least the preceding three years. See www.computer.org/election for more details on the 2007 IEEE Computer Society elections. ■ John Vig Named IEEE President-Elect for 2008 IEEE members recently selected John Vig as their president-elect for 2008.Vig is an IEEE Fellow and the recipient of the IEEE’s W.G. Cady Award and C.B. Sawyer Memorial Award, which recognize outstanding contributions in frequency control. Vig will serve one year as IEEE president-elect, participating in Board of Directors activities. He will then assume the role of president in the following year.After his term in 2009, Vig will serve as past president in 2010. In the same election, IEEE members chose 2003 Computer Society president Stephen Diamond as division VIII director-elect for 2008. Diamond, a managing director at Picosoft, served as a member of the IEEE Board of Directors in 2005 and 2006. He currently serves as the chair of the IEEE Marketing and Sales Committee. Division directors represent IEEE societies on the IEEE Board of Directors and Technical Activities Board. Division directors V and VIII are elected to represent the Computer Society membership. Diamond will act as director-elect in 2008 and as division director for 2009-2010.The division directors also serve as ex officio members of the Computer Society’s Board of Governors and Executive Committee. John Vig is a technical advisor for the US Defense Advanced Research Projects Agency. 63 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F COMPUTER SOCIETY CONNECTION Computer Recognizes Expert Reviewers O rganized activities require teamwork to succeed. Since becoming Computer’s editor in chief in January 2007, I have witnessed the great synergy between the professional editorial staff, volunteer editors, and an impressive group of reviewers whose collective wisdom and effort make Computer a great publication. First, I extend my heartfelt thanks to all the reviewers who gave their time to review submitted manuscripts, offering comments on organization and clarity, questions of accuracy, disputed definitions, and the effectiveness of visual aids, figures, or other ancillary materials. By relying on such dedicated professionals, we ensure the high quality of the peer-review process that serves as a cornerstone of any first-rate professional association publication. I encourage reviewers to continue to support the magazine and make themselves available to review for Computer in the years to come. Due to the magnitude of the workload, the editors at Computer work in a highly structured manner. Reviewers for Computer mostly work under the direction of Associate Editors in Chief Kathleen Swigger and Bill Schilit. Kathy and Bill, as volunteer leaders in the Computer Society, have contributed extensively to Computer throughout the years. I offer my gratitude to the editors who solicit or contribute Computer’s columns and departments. There will continue to be transitions as veteran editors move on and new faces take charge as I begin implementing my editorial vision and plan. Finally, thanks to the area editors and advisory panel members who have offered me support and advice through the past year. Current reviewers, please visit http://cs-ieee. manuscriptcentral.com to update the profile of your areas of expertise, if applicable. Also, please ask your colleagues to join the team by registering themselves and their areas of expertise at the same Web site. —Carl Chang, Editor in Chief A list of Computer’s expert reviewers is available on the Web at www.computer.org/reviewers2007. Task Force Becomes Technical Committee on Autonomous and Autonomic Systems The IEEE Computer Society Technical Activities Board recently voted to raise the status of the growing Task Force on Autonomous and Autonomic Systems to that of a formal Technical Committee. Roy Sterritt, of the University of Ulster in Northern Ireland, serves as chair of the TC-AAS, while Mike Hinchey, of Loyola College in Maryland and the NASA Goddard Space Flight Center, serves as its vice chair. The new technical committee will continue the task force’s work promoting interest in self-managing, -governing, and -organizing systems, including autonomic networking and communications; autonomous, self-organizing and ubiquitous systems; and autonomic, grid, organic, and pervasive computing. The TC-AAS publishes an electronic newsletter and a Letters series, which includes written versions of keynote speeches given at several relevant conferences, symposia, and workshops. These letters are 64 archived in the proceedings of the International Workshop on Engineering of Autonomic Systems (EASe) in the IEEE Computer Society Digital Library (www.computer.org/csdl). Members of the TC-AAS have organized several events, including EASe in April 2008 (www.ulster.ac. uk/ease) _____ and the IEEE International Conference on Self-Adaptive and Self-Organizing Systems, which will take place in Venice, Italy, from 20-24 October 2008 (www.saso-conference.org). Several workshops and other events will take place in conjunction with the Fifth IEEE International Conference on Autonomic Computing, which convenes in Chicago 2-6 June 2008. Membership in the TC-AAS is open to Computer Society members and nonmembers alike. To sign up, learn about upcoming activities, or view archived newsletters, visit http://tab.computer.org/aas. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Society Launches New Certification Program T he IEEE Computer Society will debut the Certified Software Development Associate program in April 2008. A volunteer panel of expert software engineers, developers, and educators created the new certification in response to industry leaders’ requests for a way to assess the skill and knowledge of individuals who are just embarking on their careers as software professionals. The CSDA certification encompasses the entire field of software development and validates knowledge of the foundations of computer science, computer engineering, and mathematics. The exam covers core software engineering principles, including software construction, software design, software testing, software requirements, and software methods. CSDA certification standards are based on The Guide to the Software Engineering Body of Knowledge (SWEBOK) and Software Engineering 2004: Curriculum Guidelines for Undergraduate Degree Programs in Software Engineering (SE2004.) VIII. Software engineering process IX. Software engineering methods X. Software quality XI. Software engineering professional practice XII. Software engineering economics XIII. Computing foundations XIV. Mathematical foundations XV. Engineering foundations CSDA BETA TEST The IEEE Computer Society will administer a beta test for the CSDA exam from 7 December 2007 to 18 January 2008. Its purpose is to validate the exam questions and provide statistical data. The CSDA beta exam is open to recent graduates and students in their final year of a baccalaureate or equivalent degree program. The exam will be given at Prometric Testing Centers at locations around the world. Applications to take the test are due by 7 December. Fees for the beta exam total $110, rising to $250 when testing opens to the public in April. Candidates who pass the beta exam will be awarded the CSDA certificate and mailed the results in early March. To learn more about CSDA certification and other IEEE Computer Society professional development programs, visit www.computer.org/csda. ■ Questions on the Certified Software Development Associate exam address the following 15 areas of expertise: I. Software requirements II. Software design III. Software construction IV. Software testing V. Software maintenance VI. Software configuration management VII. Software engineering management Editor: Bob Ward, Computer; ______________ bnward@computer.org REACH HIGHER Advancing in the IEEE Computer Society can elevate your standing in the profession. Application to Senior-grade membership recognizes ✔ ten years or more of professional expertise Nomination to Fellow-grade membership recognizes ✔ exemplary accomplishments in computer engineering GIVE YOUR CAREER A BOOST ■ UPGRADE YOUR MEMBERSHIP w w w. c o m p u t e r. o r g / j o i n / g r a d e s . h t m ____________________________________________ 65 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F CALL AND CALENDAR CALLS FOR ARTICLES FOR IEEE CS PUBLICATIONS APRIL 2008 CALLS FOR PAPERS 7-12 Apr: ICDE 2008, 24th IEEE Int’l Conf. on Data Engineering, Cancun, Mexico; www.icde2008.org LICS 2008, IEEE Logic in Computer Science Symp., 2427 June, Pittsburgh; Submissions due 7 Jan. 2008; www2.informatik.hu-berlin.de/lics/lics08/ cfp08-1.pdf ______________________________________ SCC 2008, IEEE Int’l Conf. on Services Computing, 811 July, Hawai’i; Submissions due 28 Jan. 2008; ____ http:// conferences.computer.org/scc/2008/cfp.html 8-12 Apr: MCN 2008, 2nd IEEE Workshop on MissionCritical Networking (with InfoCom), Phoenix; www. ____ criticalnet.org __________ 14-15 Apr: RAW 2008, 15th Reconfigurable ____ Architectures Workshop (with IPDPS), Miami; www. ece.lsu.edu/vaidy/raw _______________ ICWS 2008, IEEE Conf. on Web Services, 23-26 Sept., Beijing; Submissions due 7 Apr. 2008; http://conferences. computer.org/icws/2008/call-for-papers.html 14-18 Apr: IPDPS 2008, 22nd IEEE Int’l Parallel and Distributed Processing Symp., Miami; www.ipdps.org CALENDAR 15-17 Apr: InfoCom 2008, 27th IEEE Conf. on Computer Communications, Phoenix; www.ieee-infocom.org JANUARY 2008 7-10 Jan: HICSS 2008, Hawai’i Int’l Conf. on System Sciences, Waikoloa, Hawai’i; www.hicss.hawaii.edu/ hicss_41/apahome41.html ___________________ FEBRUARY 2008 18-21 Feb: WICSA 2008, Working IEEE/IFIP Conf. on Software Architecture, Vancouver, Canada; www.wicsa.net MARCH 2008 18 Apr: Hot-P2P 2008, 5th Int’l Workshop on Hot Topics in Peer-to-Peer Systems (with IPDPS), Miami; www.disi.unige.it/hotp2p/2008/index.php 18 Apr: PCGrid 2008, 2nd Workshop on Desktop Grids and Volunteer Computing Systems (with IPDPS), Miami; http://pcgrid.lri.fr 18 Apr: SSN 2008, 4th Int’l Workshop on Security in Systems and Networks (with IPDPS), Miami; www.cse. buffalo.edu/~fwu2/ssn08 __________________ 3-7 Mar: SimuTools 2008, 1st Int’l Conf. on Simulation Tools and Techniques for Communications, Networks, and Systems, Vancouver, Canada; www.simutools.org 25-28 Mar: AINA 2008, 22nd IEEE Int’l Conf. on Advanced Information Networking and Applications, Okinawa, Japan; www.aina-conference.org/2008 25-28 Mar: SOCNE 2008, 3rd IEEE Workshop on Service-Oriented Architectures in Converging Networked Environments (with AINA), Okinawa, Japan; www.c-lab.de/RLS/SOCNE08 Submission Instructions The Call and Calendar section lists conferences, symposia, and workshops that the IEEE Computer Society sponsors or cooperates in presenting. Visit www.computer.org/conferences for instructions on how to submit conference or call listings as well as a more complete listing of upcoming computer-related conferences. 66 LICS 2008 The 23rd IEEE Logic in Computer Science symposium is an annual international forum on theoretical and practical topics in computer science that relate to logic. Organizers have invited submissions on topics that include automata theory, categorical models and logics, concurrency, distributed computation, logical frameworks, and constraint programming. The symposium is sponsored by the IEEE Computer Society Technical Committee on Mathematical Foundations of Computing, in cooperation with the Association for Symbolic Logic and the European Association for Theoretical Computer Science. LICS will take place 24-27 June 2008 in Pittsburgh. Abstracts are due by 7 January 2008. Visit www2. ____ informatik.hu-berlin.de/lics/lics08 for more details on _______________________ LICS 2008, including a complete call for papers. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page MAY 2008 4-8 May: VLSI 2008, 28th IEEE VLSI Test Symp., San Diego; www.tttc-vts.org/ A BEMaGS F 24-27 June: LICS 2008, IEEE Symp. on Logic in Computer Science, Pittsburgh; www2. informatik. ______________ hu-berlin.de/lics/lics08 ________________ JULY 2008 5-7 May: ISORC 2008, 11th IEEE Int’l Symp. on Object/ Component/Service-Oriented Real-Time Distributed Computing, Orlando, Florida; http://ise.gmu.edu/isorc08 7-11 July: Services 2008, IEEE Congress on Services, Hawai’i; http://conferences.computer.org/services/2008 7-9 May: EDCC 2008, 7th European Dependable Computing Conf., Kaunas, Lithuania; http://edcc. dependability.org ____________ 8-11 July: CIT 2008, IEEE Int’l Conf. on Computer and Information Technology, Sydney, Australia; ____ http:// attend.it.uts.edu.au/cit2008 ___________________ 18-22 May: CCGrid 2008, 8th IEEE Int’l Symp. on Cluster Computing and the Grid, Lyon, France; ____ http:// ccgrid2008.ens-lyon.fr ________________ 8-11 July: SCC 2008, IEEE Int’l Conf. on Services Computing, Hawai’i; http://conferences.computer.org/ scc/2008 ______ 22-24 May: ISMVL 2008, 38th Int’l Symp. on MultipleValued Logic, Dallas; http://engr.smu.edu/ismvl08 24 May: ULSI 2008, 17th Int’l Workshop on Post-Binary ULSI Systems (with ISMVL), Dallas; http://engr.smu. edu/ismvl08 _________ JUNE 2008 23-25 June: CSF 2008, 21st IEEE Computer Security Foundations Symp. (with LICS), Pittsburgh; www.cylab. cmu.edu/CSF2008 _____________ 23-25 June: WETICE 2008, 17th IEEE Int’l Workshop on Enabling Technologies: Infrastructures for Collaborative Enterprises, Rome; www.sel.uniroma2.it/ wetice08/venue.htm ______________ 23-26 June: ICITA 2008, 5th Int’l Conf. on Information Technology and Applications, Cairns, Australia; www. ____ icita.org Events in 2007-2008 JANUARY 7-10 . . . . . . . . . . . . . . . . . . . . .HICSS 2008 FEBRUARY 18-21 . . . . . . . . . . . . . . . . . . .WICSA 2008 MARCH 3-7 . . . . . . . . . . . . . . . . . .SimuTools 2008 25-28 . . . . . . . . . . . . . . . . . . . .AINA 2008 25-28 . . . . . . . . . . . . . . . . . .SOCNE 2008 Call for Articles for Computer Computer seeks articles for a July 2008 special issue on high-assurance service-oriented architectures. The guest editors are Jing Dong from the University of Texas at Dallas, Raymond Paul from the US Department of Defense, and Liang-Jie Zhang from IBM’s T.J. Watson Research Center. Recent advances in services computing technology make it possible to register, request, discover, and supply software services online. Such loosely coupled software services form a service-oriented architecture with the support of network resources. Serviceoriented architectures have been applied in many mission-critical environments including medical, traffic control, and defense systems. These systems are required to be highly reliable, secure, available, timely, fault-tolerant, and dependable. Recently, architects have come to face new challenges in developing service-oriented systems with high assurance requirements. Computer invites papers that describe techniques, tools, or experiences related to the design, development, or assessment of practical high-assurance systems. Editors are particularly interested in submissions that address applications of service-oriented techniques. Examples of suitable topics include highassurance service compositions; service specifications for security, reliability, dependability, availability, and QoS properties; service discoveries with highassurance system requirements; and service security, trust, and privacy. The deadline for papers is 14 December. Detailed author instructions are available at www.computer. org/portal/pages/computer/mc/author.html. Send inquiries to the guest editors at ______________ jdong@utdallas.edu, raymond.paul@osd.mil, or zhanglj@us.ibm.com. _______________ ________________ 67 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page 2008 MEMBERSHIP APPLICATION A BEMaGS F IEEE Computer Society FIND THE RIGHT SOLUTION! Solve problems, learn new skills, and grow your career with the cutting edge resources of the IEEE Computer Society. www.computer.org/join Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F 2008 RATES for IEEE COMPUTER SOCIETY Membership Dues and Subscriptions Membership and periodical subscriptions are annualized to and expire on 31 December 2008. Pay full or half-year rate depending upon the date of receipt by the IEEE Computer Society as noted below. Membership Options* All prices are quoted in U.S. dollars. ´ I do not belong to the IEEE and I want to join only the Computer Society: ´ I want to join both the Computer Society and the IEEE: FULL YEAR HALF YEAR Applications received 17 Aug 07 – 29 Feb 08 Applications received 1 Mar 08– 15 Aug 08 $113.00 $57.00 $220.00 $195.00 $187.00 $180.00 $181.00 $110.00 $98.00 $94.00 $90.00 $91.00 $50.00 $25.00 Payment Information Payment required with application Membership fee I reside in the USA I reside in Canada I reside in Africa/Europe/Middle East I reside in Latin America I reside in Asia/Pacific ´ I already belong to the IEEE, and I want to $ Periodicals total $ Applicable sales tax*** join the Computer Society: Are you now or were you ever a member of the IEEE? Yes No If yes, please provide member # if known: ______________________ $ TOTAL: $ Add Periodicals** BEST VALUE! IEEE Computer Society Digital Library (online only) ARTIFICIAL INTELLIGENCE IEEE Intelligent Systems IEEE Transactions on Learning Technologies† IEEE Transactions on Pattern Analysis and Machine Intelligence BIOTECHNOLOGY IEEE/ACM Transactions on Computational Biology and Bioinformatics COMPUTATION Computing in Science & Engineering COMPUTER HARDWARE IEEE Computer Architecture Letters IEEE Micro IEEE Design & Test of Computers IEEE Transactions on Computers GRAPHICS & MULTIMEDIA IEEE Computer Graphics and Applications IEEE MultiMedia IEEE Transactions on Haptics† IEEE Transactions on Visualization and Computer Graphics HISTORY OF COMPUTING IEEE Annals of the History of Computing INTERNET & DATA TECHNOLOGIES IEEE Internet Computing IEEE Transactions on Knowledge and Data Engineering IEEE Transactions on Services Computing† IT & SECURITY IT Professional IEEE Security & Privacy IEEE Transactions on Dependable and Secure Computing MOBILE COMPUTING IEEE Pervasive Computing IEEE Transactions on Mobile Computing NETWORKING IEEE Transactions on Parallel and Distributed Systems SOFTWARE IEEE Software IEEE Transactions on Software Engineering ISSUES PER YEAR FULL YEAR HALF YEAR Applications received 16 Aug 07 – 29 Feb 08 PRINT + ONLINE Applications received 1 Mar 08 – 15 Aug 08 PRINT + ONLINE Enclosed: Check/Money Order**** Charge my: MasterCard VISA American Express Diner’s Club n/a $121 $61 6 4 $43 $39 $22 $18 12 $52 $26 4 $36 $18 6 $45 $23 4 6 6 12 $29 $41 $40 $47 $15 $21 $20 $24 6 4 2 $43 $38 $31 $22 $19 $16 6 $43 $22 4 $34 $17 6 12 4 $43 $49 $39 $22 $25 $18 6 6 4 $42 $24 $33 $21 $12 $17 Allow up to 8 weeks for application processing. Allow a minimum of 6 to 10 weeks for delivery of print periodicals. 4 12 $43 $43 $22 $22 Please complete both sides of this form. 12 $47 $24 6 12 $49 $38 $25 $19 For fastest service, apply online at www.computer.org/join Card Number Exp Date (month/year) Signature USA 0nly include 5-digit billing zip code * Member dues include $25 for a 12-month subscription to Computer. ** Periodicals purchased at member prices are for the member’s personal use only. *** Canadian residents add 14% HST or 6% GST to total. AL, AZ, CO, DC, GA, IN, KY, MD, MO, NM, and WV add sales tax to periodical subscriptions. European Union residents add VAT tax to IEEE Computer Society Digital Library subscription. **** Payable to the IEEE in U.S. dollars drawn on a U.S. bank account. Please include member name and number (if known) on your check. * Member dues include $25 for a 12-month subscription to Computer magazine. ** Periodicals purchased at member pirces are for the member’s personal use only. † Online issues only. Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Enter your name as you want it to appear on correspondence. As a key identifier in our database, circle your last/surname. Female Title First name Middle Last/Surname Home address City State/Province Postal code Country Home telephone Home facsimile This information is used by society magazines to verify their annual circulation. Please refer to the audit codes and indicate your selections in the box provided. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Preferred e-mail ´ F A. Primary line of business Date of birth (Day/Month/Year) Send mail to: BEMaGS BPA Information ´ Personal Information Male A Home address Business address 19. 20. 21. 22. 23. 24. Educational Information First professional degree completed Month/Year degree received Program major/course of study College/University State/Province Highest technical degree received Program/Course of study Country 25. 26. 27. 28. Computers Computer peripheral equipment Software Office and business machines Test, measurement, and instrumentation equipment Communications systems and equipment Navigation and guidance systems and equipment Consumer electronics/appliances Industrial equipment, controls, and systems ICs and microprocessors Semiconductors, components, sub-assemblies, materials, and supplies Aircraft, missiles, space, and ground support equipment Oceanography and support equipment Medical electronic equipment OEM incorporating electronics in their end product (not elsewhere classified) Independent and university research, test and design laboratories, and consultants (not connected with a manufacturing company) Government agencies and armed forces Companies using and/or incorporating any electronic products in their manufacturing, processing, research, or development activities Telecommunications services, and telephone (including cellular) Broadcast services (TV, cable, radio) Transportation services (airlines, railroads, etc.) Computer and communications and data processing services Power production, generation, transmission, and distribution Other commercial users of electrical, electronic equipment, and services (not elsewhere classified) Distributor (reseller, wholesaler, retailer) University, college/other education institutions, libraries Retired Others (allied to this field) _______________________________ B. Principal job function Month/Year received College/University State/Province Country ´ Business/Professional Information Title/Position Years in current position Years of practice since graduation Employer name Department/Division Street address City State/Province Postal code Country 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. General and corporate management Engineering management Project engineering management Research and development management Design engineering management — analog Design engineering management — digital Research and development engineering Design/development engineering — analog Design/development engineering — digital Hardware engineering Software design/development Computer science Science/physics/mathematics Engineering (not elsewhere classified) Marketing/sales/purchasing Consulting Education/teaching Retired Other _______________________________________________ C. Principal responsibility Office phone Office facsimile I hereby make application for Computer Society and/or IEEE membership and agree to be governed by IEEE’s Constitution, Bylaws, Statements of Policies and Procedures, and Code of Ethics. I authorize release of information related to this application to determine my qualifications for membership. Signature 1. 2. 3. 4. 5. 6. 7. 8. 9. Engineering or scientific management Management other than engineering Engineering design Engineering Software: science/management/engineering Education/teaching Consulting Retired Other _______________________________________________ Date IF8L NOTE: In order for us to process your application, you must complete and return BOTH sides of this form to the office nearest you: Asia/Pacific Office IEEE Computer Society Watanabe Bldg. 1-4-2 Minami-Aoyama Minato-ku, Tokyo 107-0062 Japan Phone: +81 3 3408 3118 Fax: +81 3 3408 3553 E-mail: __________ tokyo.ofc@computer.org Computer Publications Office IEEE Computer Society 10662 Los Vaqueros Circle P.O. Box 3014 Los Alamitos, CA 90720-1314 USA Phone: +1 800 272 6657 (USA and Canada) Phone: +1 714 821 8380 (worldwide) Fax: +1 714 821 4641 E-mail: ________ help@computer.org D. Title 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Chairman of the Board/President/CEO Owner/Partner General Manager V.P. Operations V.P. Engineering/Director Engineering Chief Engineer/Chief Scientist Engineering Manager Scientific Manager Member of Technical Staff Design Engineering Manager Design Engineer Hardware Engineer Software Engineer Computer Scientist Dean/Professor/Instructor Consultant Retired Other Professional/Technical _____________________________ Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F FEATURED TITLE FROM WILEY AND CS PRESS Software Engineering: Barry W. Boehm’s Lifetime Contributions to Software Development, Management, and Research edited by Richard W. Selby 978-0-470-14873-0 June 2007 • 832 pages Hardcover • $79.95 A Wiley-IEEE CS Press Publication To Order: North America 1-877-762-2974 Rest of the World + 44 (0) 1243 843294 This is the most authoritative archive of Barry Boehm’s contributions to software engineering. Featuring 42 reprinted articles, along with an introduction and chapter summaries to provide context, it serves as a “how-to” reference manual for software engineering best practices. It provides convenient access to Boehm’s landmark work on product development and management processes. The book concludes with an insightful look to the future by Dr. Boehm. 20 % Computer Pro m CS otio nC CH 7 od e Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX Volume 40, 2007 SUBJECT INDEX A 3D body scanning 3D Body Scanning and Healthcare Applications, P. Treleaven and J. Wells, July, pp. 28-34. 3D graphics systems How GPUs Work [How Things Work], D. Luebke and G. Humphreys, Feb., pp. 96-100. 3D Internet Generation 3D: Living in Virtual Worlds [Entertainment Computing], M. Macedonia, Oct., pp. 99-101. 3D visualization 3D Display Using Passive Optical Scatterers, S.K. Nayar and V.N. Anand, July, pp. 54-63. 3D Vision: Developing an Embedded Stereo-Vision System [Embedded Computing], J.I. Woodfill et al., May, pp. 106-108. Immersidata Analysis: Four Case Studies, C. Shahabi et al., July, pp. 45-52. Virtual Reality: How Much Immersion Is Enough?, D.A. Bowman and R.P. McMahan, July, pp. 36-43. Agile software Standards, Agility, and Engineering, F. Coallier, Sept., pp. 100-102. Analytical models The Discipline of Embedded Systems Design, T.A. Henzinger and J. Sifakis, Oct., pp. 32-40. AQUA AQUA: An Amphibious Autonomous Robot, G. Dudek et al., Jan., pp. 46-53. Artificial intelligence Automated Killers and the Computing Profession [The Profession], N. Sharkey, Nov., pp. 124, 122-123. Game Smarts [Entertainment Computing], M. van Lent, Apr., pp. 99-101. The Strangest Thing About Software, T. Menzies et al., Jan., pp. 54-60. Automated traceability Best Practices for Automated Traceability, J. Cleland-Huang et al., June, pp. 27-35. Automotive electronics system design Embedded System Design for Automotive Applications, A. Sangiovanni-Vincentelli and M. Di Natale, Oct., pp. 42-51. AUTOSAR Embedded System Design for Automotive Applications, A. Sangiovanni-Vincentelli and M. Di Natale, Oct., pp. 42-51. Avionics systems The Glass Cockpit [How Things Work], J. Knight, Oct., pp. 92-95. B Binary arithmetic Binary Arithmetic [How Things Work], N. Holmes, June, pp. 90-93. Biomolecular simulations Using FPGA Devices to Accelerate Biomolecular Simulations, S.R. Alam et al., Mar., pp. 66-73. BioTracking Swarms and Swarm Intelligence [Software Technologies], M.G. Hinchey et al., Apr., pp. 111-113. BlueFS Consumer Electronics Meets Distributed Storage [Invisible Computing], D. Peek and J. Flinn, Feb., pp. 93-95. Bohrbugs Fighting Bugs: Remove, Retry, Replicate, and Rejuvenate [Software Technologies], M. Grottke and K.S. Trivedi, Feb., pp. 107-109. 72 Broadening participation in computing Increasing the Participation of People with Disabilities in Computing Fields [Broadening Participation in Computing], S.E. Burgstahler and R.E. Ladner, May, pp. 94-97. Propagating Diversity through Active Dissemination [Broadening Participation in Computing], K.A. Siek et al., Feb., pp. 89-92. Business intelligence The Current State of Business Intelligence, H.J. Watson and B.H. Wixom, Sept., pp. 96-99. Business service stack Service Is in the Eyes of the Beholder, T. Margaria, Nov., pp. 33-37. C Cache consistency protocol Data Consistency for Cooperative Caching in Mobile Environments, J. Cao et al., Apr., pp. 60-66. Caravela environment Caravela: A Novel Stream-Based Distributed Computing Environment, S. Yamagiwa and L. Sousa, May, pp. 70-77. Cell Broadband Engine An Open Source Environment for Cell Broadband Engine System Software, M. Gschwind et al., June, pp. 37-47. Change-tolerant systems An Era of Change-Tolerant Systems [Software Technologies], S. Bohner, June, pp. 100-102. Chip multiprocessors Isolation in Commodity Multicore Processors, N. Aggarwal et al., June, pp. 49-59. An Open Source Environment for Cell Broadband Engine System Software, M. Gschwind et al., June, pp. 37-47. CIRCA system A Communication Support System for Older People with Dementia, N. Alm et al., May, pp. 35-41. Classroom Presenter Classroom Presenter: Enhancing Interactive Education with Digital Ink, R. Anderson et al., Sept., pp. 56-61. Ink, Improvisation, and Interactive Engagement: Learning with Tablets, J. Roschelle et al., Sept., pp.42-48. Click fraud Click Fraud [Web Technologies], B.J. Jansen, July, pp. 85-86. New Technology Prevents Click Fraud [News], L.D. Paulson, Mar., pp. 20-22. Client-server systems Empirical Test Observations in Client-Server Systems, L. Hatton, May, pp. 24-29. Cognitive assistance A Communication Support System for Older People with Dementia, N. Alm et al., May, pp. 35-41. Collaboration environments Supporting Resource-Constrained Collaboration Environments [The Profession], S.M. Price, June, pp. 108, 106-107. Collaborative Web search A Community-Based Approach to Personalizing Web Search, B. Smyth, Aug., pp. 42-50. Collaboratories Examining the Challenges of Scientific Workflows [Computing Practices], Y. Gil et al., Dec., pp. 24-32. Competisoft project Software Process Improvement: The Competisoft Project, Oktaba H. et al., Oct., pp. 21-28. Computational models The Discipline of Embedded Systems Design, T.A. Henzinger and J. Sifakis, Oct., pp. 32-40. Computer architectures Architectures for Silicon Nanoelectronics and Beyond, R.I. Bahar et al., Jan., pp. 25-33. The Embedded Systems Landscape [Guest Editor’s Introduction], W. Wolf, Oct., pp. 29-31. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Computer performance evaluation A New Era of Performance Evaluation [Perspectives], S.M. Pieper et al., Sept., pp. 23-30. Computers and society Annie and the Boys [In Our Time], D.A. Grier, Aug., pp. 6-9. Automated Killers and the Computing Profession [The Profession], N. Sharkey, Nov., pp. 124, 122-23. The Best Deal in Town [In Our Time], D.A. Grier, Apr., pp. 8-11. The Boundaries of Time [In Our Time], D.A. Grier, July, pp. 5-7. The Camino Real [In Our Time], D.A. Grier, June, pp. 6-8. The Chimera of Software Quality [The Profession], L. Hatton, Aug., pp. 104, 102-103. Computing as an Evolving Discipline: 10 Observations [The Profession], J. Liu, May, pp. 112, 110-111. The Computing Profession and Higher Education [The Profession], N. Holmes, Jan., pp. 116, 114-115. Consciousness and Computers [The Profession], N. Holmes, July, pp. 100, 98-99. Controlling the Conversation, [In Our Time], D.A. Grier, Sept., pp. 7-9. Counting Beans [In Our Time], D.A. Grier, Nov., pp. 8-10. Digital Technology and the Skills Shortage [The Profession], N. Holmes, Mar., pp. 100, 98-99. Dirty Electricity [In Our Time], D.A. Grier, Feb., pp. 6-8. E-Mailing from Armenia [In Our Time], D.A. Grier, Oct., pp. 8-10. A Force of Nature, [In Our Time], D.A. Grier, Dec., pp. 8-9. Getting Real in the Classroom [The Profession], M. van Genuchten and D. Vogel, Oct., pp. 108, 109-110. Incorporating a Variable-Expertise-Level System in IT Course Modules [The Profession], N. Harkiolakis, Apr., pp. 116, 114-115. Making Computers Do More with Less [The Profession], S. Santini, Dec., pp. 124, 122-123. Outposts [In Our Time], D.A. Grier, Mar., pp. 8-10. The Profession as a Culture Killer, [The Profession], N. Holmes, Sept., pp. 112, 110-11. Supporting Resource-Constrained Collaboration Environments [The Profession], S.M. Price, June, pp. 108, 106-107. The Wave of the Future [In Our Time], D.A. Grier, Jan., pp. 12-14. Working Class Hero [In Our Time], D.A. Grier, May, pp. 8-10. Computers in education Classroom Presenter: Enhancing Interactive Education with Digital Ink, R. Anderson et al., Sept., pp. 56-1. Computing as an Evolving Discipline: 10 Observations [The Profession], J. Liu, May, pp. 112, 110-111. Facilitating Pedagogical Practices through a Large-Scale Tablet PC Deployment, J.G. Tront, Sept., pp. 62-68. Getting Real in the Classroom [The Profession], M. van Genuchten and D. Vogel, Oct., pp. 108, 109-110. Handwriting Recognition: Tablet PC Text Input, J.A. Pittman, Sept., pp. 49-54. Incorporating a Variable-Expertise-evel System in IT Course Modules [The Profession], N. Harkiolakis, Apr., pp. 116, 114-115. Ink, Improvisation, and Interactive Engagement: Learning with Tablets, J. Roschelle et al., Sept., pp.42-48. Magic Paper: Sketch-Understanding Research, R. Davis, Sept., pp. 34-41. Tablet PC Technology: The Next Generation [Guest Editors’ Introduction], J. Prey and A. Weaver, Sept., pp. 32-33. Computing profession Automated Killers and the Computing Profession [The Profession], N. Sharkey, Nov., pp. 124, 122-123. The Chimera of Software Quality [The Profession], L. Hatton, Aug., pp. 104, 102-103 Computing as an Evolving Discipline: 10 Observations [The Profession], J. Liu, May, pp. 112, 110-111. The Computing Profession and Higher Education [The Profession], N. Holmes, Jan., pp. 116, 114-115. Consciousness and Computers [The Profession], N. Holmes, July, pp. 100, 98-99. Digital Technology and the Skills Shortage [The Profession], N. Holmes, Mar., pp. 100, 98-99. Getting Real in the Classroom [The Profession], M. van Genuchten and D. Vogel, Oct., pp. 108, 109-110. Incorporating a Variable-Expertise-Level System in IT Course Modules [The Profession], N. Harkiolakis, Apr., pp. 116, 114-115. A BEMaGS F Increasing the Participation of People with Disabilities in Computing Fields [Broadening Participation in Computing], S.E. Burgstahler and R.E. Ladner, May, pp. 94-97. Making Computers Do More with Less [The Profession], S. Santini, Dec., pp. 124, 122-123. The Profession as a Culture Killer, [The Profession], N. Holmes, Sept., pp. 112, 110-111. Software Development: What Is the Problem? [The Profession], R.R. Loka, Feb., pp. 112, 110-111. Supporting Resource-Constrained Collaboration Environments [The Profession], S.M. Price, June, pp. 108, 106-107. Concurrent versioning system Teaching Software Evolution in Open Source [Computing Practices], M. Petrenko et al., Nov., pp. 25-31. Consumer electronics Consumer Electronics Meets Distributed Storage [Invisible Computing], D. Peek and J. Flinn, Feb., pp. 93-95. The Future Arrives? Finally [Entertainment Computing], M. Macedonia, Feb., pp. 101-103. Cryptography Cryptography on a Speck of Dust, J.-P. Kaps et al., Feb., pp. 38-44. Resolving the Micropayment Problem [Security], M. Tripunitara and T. Messerges, Feb., pp. 104-106. Cybersecurity The Case for Flexible NIST Security Standards, F. Keblawi and D. Sullivan, June, pp. 19-26. D Database conceptual schemas Database Conceptual Schema Matching [Software Technologies], M.A. Casanova et al., Oct., pp. 102-104. Data management Data Consistency for Cooperative Caching in Mobile Environments, J. Cao et al., Apr., pp. 60-66. A Data Integration Broker for Healthcare Systems, D. Budgen et al., Apr., pp. 34-41. Immersidata Analysis: Four Case Studies, C. Shahabi et al., July, pp. 45-52. Improving Data Accessibility with File Area Networks [News], D. Geer, Nov., pp. 14-17. Measuring Data Management Practice Maturity: A Community’s Self-Assessment, P. Aiken et al., Apr., pp. 42-50. Privacy-Preserving Data Mining Systems, N. Zhang and W. Zhao, Apr., pp. 52-58. Taking a Hard-Line Approach to Encryption [News], C. Laird, Mar., pp. 13-15. Data mining Privacy-Preserving Data Mining Systems, N. Zhang and W. Zhao, Apr., pp. 52-58. Process Query Systems, G. Cybenko and V.H. Berk, Jan., pp. 62-70. The Strangest Thing About Software, T. Menzies et al., Jan., pp. 54-60. Data storage Discryption: Internal Hard-Disk Encryption for Secure Storage [Security], L. Hars, June, pp. 103-105. Supporting Resource-Constrained Collaboration Environments [The Profession], S.M. Price, June, pp. 108, 106-107. DaVinci technology Using DaVinci Technology for Digital Video Devices, D. Talla and J. Golston, Oct., pp. 53-61. Debugging Boosting Debugging Support for Complex Systems on Chip, A. Mayer et al., Apr., pp. 76-81. Dirty Electricity [In Our Time], D.A. Grier, Feb., pp. 6-8. Fighting Bugs: Remove, Retry, Replicate, and Rejuvenate [Software Technologies], M. Grottke and K.S. Trivedi, Feb., pp. 107-109. Denial-of-service attacks Marking Technique to Isolate Boundary Router and Attacker, V. Vijairaghavan et al., Feb., pp. 54-58. Dictionary attacks Password-Based Authentication: Preventing Dictionary Attacks, S. Chakrabarti and M. Singhal, June, pp. 68-74. Digital ink Classroom Presenter: Enhancing Interactive Education with Digital Ink, R. Anderson et al., Sept., pp. 56-61. 73 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX Digital technology Consciousness and Computers [The Profession], N. Holmes, July, pp. 100, 98-99. The Future Arrives? Finally [Entertainment Computing], M. Macedonia, Feb., pp. 101-103. Researchers Develop Efficient Digital-Cameras Design [News], L.D. Paulson, Mar., pp. 20-22. Using DaVinci Technology for Digital Video Devices, D. Talla and J. Golston, Oct., pp. 53-61. Discryption Discryption: Internal Hard-Disk Encryption for Secure Storage [Security], L. Hars, June, pp. 103-105. Distributed computing Caravela: A Novel Stream-Based Distributed Computing Environment, S. Yamagiwa and L. Sousa, May, pp. 70-77. Consumer Electronics Meets Distributed Storage [Invisible Computing], D. Peek and J. Flinn, Feb., pp. 93-95. A Data Integration Broker for Healthcare Systems, D. Budgen et al., Apr., pp. 34-41. Trust Management in Distributed Systems, H. Li and M. Singhal, Feb., pp. 45-53. Entertainment computing Enhancing the User Experience in Mobile Phones [Entertainment Computing], S.R. Subramanya and B.K. Yi, Dec., pp. 114-117. The Future Arrives? Finally [Entertainment Computing], M. Macedonia, Feb., pp. 101-103. Game Smarts [Entertainment Computing], M. van Lent, Apr., pp. 99-101. Games: Once More, with Feeling [Entertainment Computing], M. van Lent and W. Swartout, Aug., pp. 98-100. Generation 3D: Living in Virtual Worlds [Entertainment Computing], M. Macedonia, Oct., pp. 99-101. iPhones Target the Tech Elite [Entertainment Computing], M. Macedonia, June, pp. 94-95. Environmental monitoring Process Query Systems, G. Cybenko and V.H. Berk, Jan., pp. 62-70. Escher model Escher: A New Technology Transitioning Model [Embedded Computing], J. Sztipanovits et al., Mar., pp. 90-92. E Fault isolation Isolation in Commodity Multicore Processors, N. Aggarwal et al., June, pp. 49-59. FPGAs Achieving High Performance with FPGA-Based Computing, M.C. Herbordt et al., Mar., pp. 50-57. High-Performance Reconfigurable Computing [Guest Editors’ Introduction], D. Buell et al., Mar., pp. 23-27. It’s Time to Stop Calling Circuits “Hardware” [Embedded Computing], F. Vahid, Sept., pp. 106-108. Sparse Matrix Computations on Reconfigurable Hardware, V.K. Prasanna and G.R. Morris, Mar., pp. 58-64. Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. Using FPGA Devices to Accelerate Biomolecular Simulations, S.R. Alam et al., Mar., pp. 66-73. Flow-model-based computation Caravela: A Novel Stream-Based Distributed Computing Environment, S. Yamagiwa and L. Sousa, May, pp. 70-77. Formal methods Designing for Software’s Social Complexity, J.L. Fiadeiro, Jan., pp. 34-39. FT-CORBA Fault-Tolerant CORBA: From Specification to Reality [Standards], P. Narasimhan, Jan., pp. 110-112. E-government New Paradigms for Next-Generation E-Government Projects, S. Friedrichs and S. Jung, Nov., pp. 53-55. E-mail management Managing E-Mail Overload: Solutions and Future Challenges, D. Schuff et al., Feb., pp. 31-36. E-passports Replacing Lost or Stolen E-Passports [Security], J. Yong and E. Bertino, Oct., pp. 89-91. Ecosystem integration Toward Mobile Services: Three Approaches, J. Bosch, Nov., pp. 51-53. Electronic voting Electronic Voting [How Things Work], J. Epstein, Aug., pp. 92-95. Embedded computing 3D Vision: Developing an Embedded StereoVision System [Embedded Computing], J.I. Woodfill et al., May, pp. 106-108. Boosting Debugging Support for Complex Systems on Chip, A. Mayer et al., Apr., pp. 76-81. The Discipline of Embedded Systems Design, T.A. Henzinger and J. Sifakis, Oct., pp. 32-40. Embedded System Design for Automotive Applications, A. Sangiovanni-Vincentelli and M. Di Natale, Oct., pp. 42-51. The Embedded Systems Landscape [Guest Editor’s Introduction], Wolf W., Oct., pp. 29-31. Escher: A New Technology Transitioning Model [Embedded Computing], J. Sztipanovits et al., Mar., pp. 90-92. The Good News and the Bad News [Embedded Computing], W. Wolf, Nov., pp. 104-105. It’s Time to Stop Calling Circuits “Hardware” [Embedded Computing], F. Vahid, Sept., pp. 106-108. SensorMap for Wide-Area Sensor Webs [Embedded Computing], S. Nath et al., July, pp. 90-93. Software-Defined Radio Prospects for Multistandard Mobile Phones, U. Ramacher, Oct., pp. 62-69. Using DaVinci Technology for Digital Video Devices, D. Talla and J. Golston, Oct., pp. 53-61. Energy-efficiency optimizations The Case for Energy-Proportional Computing, L.A. Barroso and U. Hölzle, Dec., pp. 33-37. Models and Metrics to Enable Energy-Efficiency Optimizations, S. Rivoire et al., Dec., pp. 39-48. Enterprise engineering Enterprise, Systems, and Software Engineering—The Need for Integration [Standards], P. Joannou, May, pp. 103-105. Enterprise services Enterprise Security for Web 2.0 [IT Systems Perspectives], M.A. Davidson, and E. Yoran, Nov., pp. 117-119. Getting on Board the Enterprise Service Bus [News], S. Ortiz Jr., Apr., pp. 15-17. Toward the Realization of Policy-Oriented Enterprise Management, M. Kaiser, Nov., pp. 57-63. Service Is in the Eyes of the Beholder, T. Margaria, Nov., pp. 33-37. Service-Oriented Computing: State of the Art and Research Challenges, M.P. Papazoglou et al., Nov., pp. 38-45. 74 F G Game technology Game Smarts [Entertainment Computing], M. van Lent, Apr., pp. 99-101. Game-theoretic analysis Reengineering the Internet for Better Security, M. Parameswaran et al., Jan., pp. 40-44. GeForce 3 How GPUs Work [How Things Work], D. Luebke and G. Humphreys, Feb., pp. 96-100. GPUs How GPUs Work [How Things Work], D. Luebke and G. Humphreys, Feb., pp. 96-100. Graphical tools The Inevitable Cycle: Graphical Tools and Programming Paradigms, J. Soukup and M. Soukup, Aug., pp. 24-30. Graphics processing units Caravela: A Novel Stream-Based Distributed Computing Environment, S. Yamagiwa and L. Sousa, May, pp. 70-77. Green Computing The Case for Energy-Proportional Computing, L.A. Barroso and U. Hölzle, Dec., pp. 33-37. The Green500 List: Encouraging Sustainable Supercomputing, W. C. Feng and K. Cameron, Dec., pp. 50-55. Life Cycle Aware Computing: Reusing Silicon Technology, J.Y. Oliver et al., Dec., pp. 56-61. Models and Metrics to Enable Energy-Efficiency Optimizations, S. Rivoire et al., Dec., pp. 39-48. Green Destiny The Green500 List: Encouraging Sustainable Supercomputing, W. C. Feng and K. Cameron, Dec., pp. 50-55. Group Scribbles Ink, Improvisation, and Interactive Engagement: Learning with Tablets, J. Roschelle et al., Sept., pp.42-48. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page H Handel-C Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. Handwriting recognition Handwriting Recognition: Tablet PC Text Input, J.A. Pittman, Sept., pp. 49-54. Hard-disk encryption Discryption: Internal Hard-Disk Encryption for Secure Storage [Security], L. Hars, June, pp. 103-105. Heads-up display Holistic Sensing and Active Displays for Intelligent Driver Support Systems, M.M. Trivedi and S.Y. Cheng, May, pp. 60-68. Healthcare technology 3D Body Scanning and Healthcare Applications, P. Treleaven and J. Wells, July, pp. 28-34. A Data Integration Broker for Healthcare Systems, D. Budgen et al., Apr., pp. 34-41. Immersidata Analysis: Four Case Studies, C. Shahabi et al., July, pp. 45-52. How Things Work Binary Arithmetic [How Things Work], N. Holmes, June, pp. 90-93. Electronic Voting [How Things Work], J. Epstein, Aug., pp. 92-95. The Glass Cockpit [How Things Work], J. Knight, Oct., pp. 92-95. How GPUs Work [How Things Work], D. Luebke and G. Humphreys, Feb., pp. 96-100. SMS: The Short Message Service [How Things Work], J. Brown et al., Dec., pp. 106-110. Wi-Fi—The Nimble Musician in Your Laptop [How Things Work], D.G. Leeper, Apr., pp. 108-110. Human activity language A Language for Human Action, G. Guerra-Filho and Y. Aloimonos, May, pp. 42-51. Human-centered computing A Communication Support System for Older People with Dementia, N. Alm et al., May, pp. 35-41. Holistic Sensing and Active Displays for Intelligent Driver Support Systems, M.M. Trivedi and S.Y. Cheng, May, pp. 60-68. Human-Centered Computing: Toward a Human Revolution [Guest Editors’ Introduction], A. Jaimes et al., May, pp. 3034. An Interactive Multimedia Diary for the Home, G.C. de Silva et al., May, pp. 52-59. A Language for Human Action, G. Guerra-Filho and Y. Aloimonos, May, pp. 42-51. I IEEE 1667 Authentication in Transient Storage Device Attachments [Security], D. Rich, Apr., pp. 102-104. IEEE Computer Society CHC61 Sites Highlight “Unsung Heroes” in 2007 [CS Connection], B. Ward, Mar., pp. 77-79. Computer Recognizes Expert Reviewers [CS Connection], C.K. Chang, Dec., pp. 64-65. Computer Society Announces Larson, UPE, and OCA Student Winners [CS Connection], B. Ward, Apr., pp. 89-91. Computer Society and IEEE Foundation Offer Cash Prizes at Intel Science Fair [CS Connection], B. Ward, Mar., pp. 7779. Computer Society Launches New Certification Program [CS Connection], B. Ward, Dec., pp. 64-65. Computer Society Recognizes Outstanding Professionals, [CS Connection], B. Ward, May, pp. 85-88. Computer Society Summer and Fall Conferences, [CS Connection], B. Ward, May, pp. 85-88. Edward Seidel Honored with Sidney Fernbach Award [CS Connection], B. Ward, Apr., pp. 89-91. Hosaka and Spielberg Named Winners of 2006 Computer Pioneer Award [CS Connection], B. Ward, Feb., pp. 73-77. IEEE Computer Society Elections, Sept., pp. 69-76. An Interesting Year [President’s Message], M.R. Williams, Dec., pp. 6-7. James Pomerene Garners Joint IEEE/ACM Award [CS Connection], B. Ward, Apr., pp. 89-91. My Vision for Computer [EIC’s Message], C.K. Chang, Jan., pp. 7-8. A BEMaGS F Tadashi Watanabe Receives 2006 Seymour Cray Award [CS Connection], B. Ward, Apr., pp. 89-91. A Year of Decision [President’s Message], M.R. Williams, Jan., pp. 9-11. Image retrieval systems From Pixels to Semantic Spaces: Advances in Content-Based Image Retrieval, N. Vasconcelos, July, pp. 20-26. Image wall The Shannon Portal Installation: Interaction Design for Public Places, L. Ciolfi et al., July, pp. 64-71. Immersidata Immersidata Analysis: Four Case Studies, C. Shahabi et al., July, pp. 45-52. iMouse iMouse: An Integrated Mobile Surveillance and Wireless Sensor System, Y.-C. Tseng et al., June, pp. 60-66. Impulse C Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. In situ computing Five Enablers for Mobile 2.0 [Invisible Computing], W.G. Griswold, Oct., pp. 96-98. Information overload An Information Avalanche [Web Technologies], V.G. Cerf, Jan., pp. 104-105. Managing E-Mail Overload: Solutions and Future Challenges, D. Schuff et al., Feb., pp. 31-36. Information storage and retrieval Measuring Data Management Practice Maturity: A Community’s Self-Assessment, P. Aiken et al., Apr., pp. 42-50. Privacy-Preserving Data Mining Systems, N. Zhang and W. Zhao, Apr., pp. 52-58. Instruction set architectures Embracing and Extending 20th-Century Instruction Set Architectures, J. Gebis and D. Patterson, Apr., pp. 68-75. Intelligent Network Evolution of SOA Concepts in Telecommunications, T. Magedanz, et al., Nov., pp. 4650. Service Is in the Eyes of the Beholder, T. Margaria, Nov., pp. 33-37. Interactive displays The Shannon Portal Installation: Interaction Design for Public Places, L. Ciolfi et al., July, pp. 64-71. Internet An Information Avalanche [Web Technologies], V.G. Cerf, Jan., pp. 104-105. Reengineering the Internet for Better Security, M. Parameswaran et al., Jan., pp. 40-44. These Are Not Your Father’s Widgets [News], G. Lawton, July, pp. 10-13. Internet communities Social Scripting for the Web [Invisible Computing], T. Lau, June, pp. 96-98. Internet Protocol traceback Marking Technique to Isolate Boundary Router and Attacker, V. Vijairaghavan et al., Feb., pp. 54-58. Intrusion detection Natural-Language Processing for Intrusion Detection [Security], A. Stone, Dec., pp. 103-105. Invisible computing Consumer Electronics Meets Distributed Storage [Invisible Computing], D. Peek and J. Flinn, Feb., pp. 93-95. Five Enablers for Mobile 2.0 [Invisible Computing], W.G. Griswold, Oct., pp. 96-98. How-To Web Pages [Invisible Computing], C. Torrey and D.W. McDonald, Aug., pp. 96-97. Predestination: Where Do You Want to Go Today? [Invisible Computing], J. Krumm and E. Horvitz, Apr., pp. 105-107. Social Scripting for the Web [Invisible Computing], T. Lau, June, pp. 96-98. iPhone iPhones Target the Tech Elite [Entertainment Computing], M. Macedonia, June, pp. 94-95. IP Multimedia Subsystem Evolution of SOA Concepts in Telecommunications, T. Magedanz et al., Nov., pp. 46-50. IP protection An Information Avalanche [Web Technologies], V.G. Cerf, Jan., pp. 104-105. 75 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX IT architecture Creating Business Value through Flexible IT Architecture, J. Helbig and A. Scherdin, Nov., pp. 55-56. IT education Incorporating a Variable-Expertise-Level System in IT Course Modules [The Profession], N. Harkiolakis, Apr., pp. 116, 114-115. IT systems perspectives The Current State of Business Intelligence, H.J. Watson and B.H. Wixom, Sept., pp. 96-99. Enterprise Security for Web 2.0 [IT Systems Perspectives], M.A. Davidson, and E. Yoran, Nov., pp. 117-119. The Impact of Software Growth on the Electronics Industry [IT Systems Perspectives], M. van Genuchten, Jan., pp. 106108. IT Audit: A Critical Business Process [IT Systems Perspectives], A. Carlin and F. Gallegos, July, pp. 87-89. Replacing Proprietary Software on the Desktop [IT Systems Perspectives], D. Hardaway, Mar., pp. 96-97. Service Management: Driving the Future of IT [IT Systems Perspectives], B. Clacy and B. Jennings, May, pp. 98-100. J J2ME technology Challenges in Securing Networked J2ME Applications, A.N. Klingsheim et al., Feb., pp. 24-30. Jacobi method Sparse Matrix Computations on Reconfigurable Hardware, V.K. Prasanna and G.R. Morris, Mar., pp. 58-64. Java Application Building Center (jABC) Full Life-Cycle Support for End-to-End Processes, B. Steffen and P. Narayan, Nov., pp. 64-73. JouleSort benchmark Models and Metrics to Enable Energy-Efficiency Optimizations, S. Rivoire et al., Dec., pp. 39-48. K Knowledge management Managing E-Mail Overload: Solutions and Future Challenges, D. Schuff et al., Feb., pp. 31-36. Koala Social Scripting for the Web [Invisible Computing], T. Lau, June, pp. 96-98. L Life cycle aware computing Life Cycle Aware Computing: Reusing Silicon Technology, J.Y. Oliver et al., Dec., pp. 56-61. Linguistics A Language for Human Action, G. Guerra-Filho and Y. Aloimonos, May, pp. 42-51. Location-based services Predestination: Where Do You Want to Go Today? [Invisible Computing], J. Krumm and E. Horvitz, Apr., pp. 105-107. LURCH The Strangest Thing About Software, T. Menzies et al., Jan., pp. 54-60. M Machine learning Search Engines that Learn from Implicit Feedback, T. Joachims and F. Radlinski, Aug., pp. 34-40. Magic paper Magic Paper: Sketch-Understanding Research, R. Davis, Sept., pp. 34-41. Malware News Briefs, L.D. Paulson, July, pp. 17-19. Reengineering the Internet for Better Security, M. Parameswaran et al., Jan., pp. 40-44. Mandelbugs Fighting Bugs: Remove, Retry, Replicate, and Rejuvenate [Software Technologies], M. Grottke and K.S. Trivedi, Feb., pp. 107-109. Metadata Measuring Data Management Practice Maturity: A Community’s Self-Assessment, P. Aiken et al., Apr., pp. 42-50. Micropayments Resolving the Micropayment Problem [Security], M. Tripunitara and T. Messerges, Feb., pp. 104-106. Mitrion-C Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. 76 MoProSoft model Software Process Improvement: The Competisoft Project, Oktaba H. et al., Oct., pp. 21-28. Mobile computing 4G Wireless Begins to Take Shape [News], S. Ortiz Jr., Nov., pp. 18-21. Adaptive QoS for Mobile Web Services through Cross-Layer Communication, M. Tian et al., Feb., pp. 59-63. Company Develops Handheld with Flexible Screen [News], L.D. Paulson, Apr., pp. 21-23. Data Consistency for Cooperative Caching in Mobile Environments, J. Cao et al., Apr., pp. 60-66. Enhancing the User Experience in Mobile Phones [Entertainment Computing], S.R. Subramanya and B.K. Yi, Dec., pp. 114-117. Five Enablers for Mobile 2.0 [Invisible Computing], W.G. Griswold, Oct., pp. 96-98. iMouse: An Integrated Mobile Surveillance and Wireless Sensor System, Y.-C. Tseng et al., June, pp. 60-66. iPhones Target the Tech Elite [Entertainment Computing], M. Macedonia, June, pp. 94-95. New Interfaces at the Touch of a Fingertip [News], S.J. Vaughan-Nichols, Aug., pp. 12-15. A New Virtual Private Network for Today’s Mobile World, [Technology News], K. Heyman, Dec., pp. 17-19. Predestination: Where Do You Want to Go Today? [Invisible Computing], J. Krumm and E. Horvitz, Apr., pp. 105-107. Software-Defined Radio Prospects for Multistandard Mobile Phones, U. Ramacher, Oct., pp. 62-69. Toward Mobile Services: Three Approaches, J. Bosch, Nov., pp. 51-53. Mobile search Deciphering Trends in Mobile Search, M. Kamvar and S. Baluja, Aug., pp. 58-62. Model-based design Embedded System Design for Automotive Applications, A. Sangiovanni-Vincentelli and M. Di Natale, Oct., pp. 42-51. Full Life-Cycle Support for End-to-End Processes, B. Steffen and P. Narayan, Nov., pp. 64-73. MuSIC-1 chip Software-Defined Radio Prospects for Multistandard Mobile Phones, U. Ramacher, Oct., pp. 62-69. Multicore processors For Programmers, Multicore Chips Mean Multiple Challenges [News], D. Geer, Sept., pp. 17-19. Isolation in Commodity Multicore Processors, N. Aggarwal et al., June, pp. 49-59. An Open Source Environment for Cell Broadband Engine System Software, M. Gschwind et al., June, pp. 37-47. Multimedia data An Interactive Multimedia Diary for the Home, G.C. de Silva et al., May, pp. 52-59. Multiprocessing The Good News and the Bad News [Embedded Computing], W. Wolf, Nov., pp. 104-105. N NIST Standards Managing Enterprise Security Risk with NIST Standards [Security], R. Ross, Aug., pp. 88-91. Nanotechnology Architectures for Silicon Nanoelectronics and Beyond, R.I. Bahar et al., Jan., pp. 25-33. Nanodevice Increases Optical Networks’ Bandwidth [News], L.D. Paulson, Apr., pp. 21-23. Natural-language processing Natural-Language Processing for Intrusion Detection [Security], A. Stone, Dec., pp. 103-105. NetBeans IDE Full Life-Cycle Support for End-to-End Processes, B. Steffen and P. Narayan, Nov., pp. 64-73. Network security Cryptography on a Speck of Dust, J.-P. Kaps et al., Feb., pp. 38-44. Marking Technique to Isolate Boundary Router and Attacker, V. Vijairaghavan et al., Feb., pp. 54-58. Trust Management in Distributed Systems, H. Li and M. Singhal, Feb., pp. 45-53. Neural networks Handwriting Recognition: Tablet PC Text Input, J.A. Pittman, Sept., pp. 49-54. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page O Online communities Online Experiments: Lessons Learned, R. Kohavi and R. Longbotham, Sept., pp. 103-105. Toward a PeopleWeb, R. Ramakrishnan and A. Tomkins, Aug., pp. 63-72. Online maps Taking Online Maps Down to Street Level [Invisible Computing], L. Vincent, Dec., pp. 118-120. Open source software The Economic Motivation of Open Source Software: Stakeholder Perspectives [News], D. Riehle, Apr., pp. 25-32. Group Works on Open Source Interoperability [News], L.D. Paulson and G. Lawton , June, pp. 15-18. An Open Source Environment for Cell Broadband Engine System Software, M. Gschwind et al., June, pp. 37-47. Replacing Proprietary Software on the Desktop [IT Systems Perspectives], D. Hardaway, Mar., pp. 96-97. Teaching Software Evolution in Open Source [Computing Practices], M. Petrenko et al., Nov., pp. 25-31. Working Class Hero [In Our Time], D.A. Grier, May, pp. 8-10. Osmot engine Search Engines that Learn from Implicit Feedback, T. Joachims and F. Radlinski, Aug., pp. 34-40. Outsourcing The Changing World of Outsourcing, [Industry Trends], N. Leavitt, Dec., pp. 13-16. P POEM Toward the Realization of Policy-Oriented Enterprise Management, M. Kaiser, Nov., pp. 57-63. Processor reuse Life Cycle Aware Computing: Reusing Silicon Technology, J.Y. Oliver et al., Dec., pp. 56-61. PQS modeling framework Process Query Systems, G. Cybenko and V.H. Berk, Jan., pp. 62-70. Pairwise preferences Search Engines that Learn from Implicit Feedback, T. Joachims and F. Radlinski, Aug., pp. 34-40. Passive optical scatterers 3D Display Using Passive Optical Scatterers, S.K. Nayar and V.N. Anand, July, pp. 54-63. Password-based authentication Password-Based Authentication: Preventing Dictionary Attacks, S. Chakrabarti and M. Singhal, June, pp. 68-74. PeopleWeb Toward a PeopleWeb, R. Ramakrishnan and A. Tomkins, Aug., pp. 63-72. Perrow-class failures Conquering Complexity [Software Technologies] G.J. Holzmann, Dec., pp. 111-113. Personalized Web search A Community-Based Approach to Personalizing Web Search, B. Smyth, Aug., pp. 42-50. Pervasive computing Cryptography on a Speck of Dust, J.-P. Kaps et al., Feb., pp. 38-44. PowerPC Embracing and Extending 20th-Century Instruction Set Architectures, J. Gebis and D. Patterson, Apr., pp. 68-75. Privacy Privacy-Preserving Data Mining Systems, N. Zhang and W. Zhao, Apr., pp. 52-58. Process detection Process Query Systems, G. Cybenko and V.H. Berk, Jan., pp. 62-70. Programming paradigms The Inevitable Cycle: Graphical Tools and Programming Paradigms, J. Soukup and M. Soukup, Aug., pp. 24-30. Public-key cryptosystems Cryptography on a Speck of Dust, J.-P. Kaps et al., Feb., pp. 38-44. R Reconfigurable computing Achieving High Performance with FPGA-Based Computing, M.C. Herbordt et al., Mar., pp. 50-57. High-Performance Reconfigurable Computing [Guest Editors’ Introduction], D. Buell et al., Mar., pp. 23-27. Software-Defined Radio Prospects for Multistandard Mobile Phones, U. Ramacher, Oct., pp. 62-69. A BEMaGS F Sparse Matrix Computations on Reconfigurable Hardware, V.K. Prasanna and G.R. Morris, Mar., pp. 58-64. Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. Using FPGA Devices to Accelerate Biomolecular Simulations, S.R. Alam et al., Mar., pp. 66-73. Vforce: An Extensible Framework for Reconfigurable Supercomputing, N. Moore et al., Mar., pp. 39-49. Robotics AQUA: An Amphibious Autonomous Robot, G. Dudek et al., Jan., pp. 46-53. Automated Killers and the Computing Profession [The Profession], N. Sharkey, Nov., pp. 124, 122-123. S Safety-critical software Unsafe Standardization [Standards], M. Thomas, Nov., pp. 109-111. SAP enterprise SOA Toward the Realization of Policy-Oriented Enterprise Management, M. Kaiser, Nov., pp. 57-63. Scanning technologies 3D Body Scanning and Healthcare Applications, P. Treleaven and J. Wells, July, pp. 28-34. Scenario-oriented computing A New Era of Performance Evaluation [Perspectives], S.M. Pieper et al., Sept., pp. 23-30. Schema matching Database Conceptual Schema Matching [Software Technologies], M.A. Casanova, Oct., pp. 102-104. Scientific workflows Examining the Challenges of Scientific Workflows [Computing Practices], Y. Gil et al., Dec., pp. 24-32. SDR baseband processors Software-Defined Radio Prospects for Multistandard Mobile Phones, U. Ramacher, Oct., pp. 62-69. Search A Community-Based Approach to Personalizing Web Search, B. Smyth, Aug., pp. 42-50. Deciphering Trends in Mobile Search, M. Kamvar and S. Baluja, Aug., pp. 58-62. Enterprise Security for Web 2.0 [IT Systems Perspectives], M.A. Davidson, and E. Yoran, Nov., pp. 117-119. Search Engines that Learn from Implicit Feedback, T. Joachims and F. Radlinski, Aug., pp. 34-40. Search: The New Incarnations [From the Area Editor], N. Ramakrishnan, Aug., pp. 31-32. Sponsored Search: Is Money a Motivator for Providing Relevant Results?, B.J. Jansen and A. Spink, Aug., pp. 52-57. Toward a PeopleWeb, R. Ramakrishnan and A. Tomkins, Aug., pp. 63-72. Security Authentication in Transient Storage Device Attachments [Security], D. Rich, Apr., pp. 102-104. The Case for Flexible NIST Security Standards, F. Keblawi and D. Sullivan, June, pp. 19-26. Challenges in Securing Networked J2ME Applications, A.N. Klingsheim et al., Feb., pp. 24-30. Discryption: Internal Hard-Disk Encryption for Secure Storage [Security], L. Hars, June, pp. 103-105. IT Audit: A Critical Business Process [IT Systems Perspectives], A. Carlin and F. Gallegos, July, pp. 87-89. Managing Enterprise Security Risk with NIST Standards [Security], R. Ross, Aug., pp. 88-91. Natural-Language Processing for Intrusion Detection [Security], A. Stone, Dec., pp. 103-105. Password-Based Authentication: Preventing Dictionary Attacks, S. Chakrabarti and M. Singhal, June, pp. 68-74. Process Query Systems, G. Cybenko and V.H. Berk, Jan., pp. 62-70. Protecting Networks by Controlling Access [News], S. Ortiz Jr., Aug., pp. 16-19. Reengineering the Internet for Better Security, M. Parameswaran et al., Jan., pp. 40-44. Replacing Lost or Stolen E-Passports [Security], J. Yong and E. Bertino, Oct., pp. 89-91. Resolving the Micropayment Problem [Security], M. Tripunitara and T. Messerges, Feb., pp. 104-106. Stronger Domain Name System Thwarts Root-Server Attacks [News], G. Lawton, May, pp. 14-17. 77 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX Web 2.0 Creates Security Challenges [News], G. Lawton, Oct., pp. 13-16. Semantic Web Toward a Social Semantic Web [Web Technologies], A. Mikroyannidis, Nov., pp. 113-115. SensorMap SensorMap for Wide-Area Sensor Webs [Embedded Computing], S. Nath et al., July, pp. 90-93. Service orientation Component Contracts in Service-Oriented Architectures, F. Curbera, Nov., pp. 74-80. Creating Business Value through Flexible IT Architecture, J. Helbig and A. Scherdin, Nov., pp. 55-56. An Era of Change-Tolerant Systems [Software Technologies], S. Bohner, June, pp. 100-102. Evolution of SOA Concepts in Telecommunications, T. Magedanz, et al., Nov., pp. 46-50. The Fractal Nature of Web Services [Web Technologies], C. Bussler, Mar., pp. 93-95. Full Life-Cycle Support for End-to-End Processes, B. Steffen and P. Narayan, Nov., pp. 64-73. Getting on Board the Enterprise Service Bus [News], S. Ortiz Jr., Apr., pp. 15-17. New Paradigms for Next-Generation E-Government Projects, S. Friedrichs and S. Jung, Nov., pp. 53-55. Service Is in the Eyes of the Beholder, T. Margaria, Nov., pp. 33-37. Service-Oriented Computing: State of the Art and Research Challenges, M.P. Papazoglou et al., Nov., pp. 38-45. Steps Toward a Science of Service Systems, J. Spohrer et al., Jan., pp. 71-77. Toward Mobile Services: Three Approaches, J. Bosch, Nov., pp. 51-53. Toward the Realization of Policy-Oriented Enterprise Management, M. Kaiser, Nov., pp. 57-63. Service provider certification Reengineering the Internet for Better Security, M. Parameswaran et al., Jan., pp. 40-44. Silicon devices Architectures for Silicon Nanoelectronics and Beyond, R.I. Bahar et al. , Jan., pp. 25-33. SIMD processors Embracing and Extending 20th-Century Instruction Set Architectures, J. Gebis and D. Patterson, Apr., pp. 68-75. SIMNET Generation 3D: Living in Virtual Worlds [Entertainment Computing], M. Macedonia, Oct., pp. 99-101. Situational awareness systems Holistic Sensing and Active Displays for Intelligent Driver Support Systems, M.M. Trivedi and S.Y. Cheng, May, pp. 60-68. Sketch-understanding systems Magic Paper: Sketch-Understanding Research, R. Davis, Sept., pp. 34-41. Smart home An Interactive Multimedia Diary for the Home, G.C. de Silva et al., May, pp. 52-59. SMS SMS: The Short Message Service [How Things Work], J. Brown et al., Dec., pp. 106-110. SOAs Component Contracts in Service-Oriented Architectures, F. Curbera, Nov., pp. 74-80. SOC paradigm Service-Oriented Computing: State of the Art and Research Challenges, M.P. Papazoglou et al., Nov., pp. 38-45. SoCs Boosting Debugging Support for Complex Systems on Chip, A. Mayer et al., Apr., pp. 76-81. Social Web Social Scripting for the Web [Invisible Computing], T. Lau, June, pp. 96-98. Toward a Social Semantic Web [Web Technologies], A. Mikroyannidis, Nov., pp. 113-115. Software-defined radio technologies Software-Defined Radio Prospects for Multistandard Mobile Phones, U. Ramacher, Oct., pp. 62-69. Software development The Chimera of Software Quality [The Profession], L. Hatton, Aug., pp. 104, 102-103. 78 A Data Integration Broker for Healthcare Systems, D. Budgen et al., Apr., pp. 34-41. The Economic Motivation of Open Source Software: Stakeholder Perspectives [News], D. Riehle, Apr., pp. 2532. Empirical Test Observations in Client-Server Systems, L. Hatton, May, pp. 24-29. Fighting Bugs: Remove, Retry, Replicate, and Rejuvenate [Software Technologies], M. Grottke and K.S. Trivedi, Feb., pp. 107-109. How Accurately Do Engineers Predict Software Maintenance Tasks?, L. Hatton, Feb., pp. 64-69. The Inevitable Cycle: Graphical Tools and Programming Paradigms, J. Soukup and M. Soukup, Aug., pp. 24-30. Software Development: What Is the Problem? [The Profession], R.R. Loka, Feb., pp. 112, 110-111. Software Process Improvement: The Competisoft Project, Oktaba H. et al., Oct., pp. 21-28. The Strangest Thing About Software, T. Menzies et al., Jan., pp. 54-60. Teaching Software Evolution in Open Source [Computing Practices], M. Petrenko et al., Nov., pp. 25-31. Unsafe Standardization [Standards], M. Thomas, Nov., pp. 109-111. Software engineering Designing for Software’s Social Complexity, J.L. Fiadeiro, Jan., pp. 34-39. The Impact of Software Growth on the Electronics Industry [IT Systems Perspectives], M. van Genuchten, Jan., pp. 106108. Software technologies Conquering Complexity [Software Technologies] G.J. Holzmann, Dec., pp. 111-113. Database Conceptual Schema Matching [Software Technologies], M.A. Casanova et al., Oct., pp. 102-104. An Era of Change-Tolerant Systems [Software Technologies], S. Bohner, June, pp. 100-102. Fighting Bugs: Remove, Retry, Replicate, and Rejuvenate [Software Technologies], M. Grottke and K.S. Trivedi, Feb., pp. 107-109. How Business Goals Drive Architectural Design [Software Technologies], R.S. Sangwan and C.J. Neill, Aug., pp. 8587. Swarms and Swarm Intelligence [Software Technologies], M.G. Hinchey et al., Apr., pp. 111-113. Sponsored search Click Fraud [Web Technologies], B.J. Jansen, July, pp. 85-86. Sponsored Search: Is Money a Motivator for Providing Relevant Results?, B.J. Jansen and A. Spink, Aug., pp. 52-57. Squirrel system Five Enablers for Mobile 2.0 [Invisible Computing], W.G. Griswold, Oct., pp. 96-98. SRC Carte Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. Standards Authentication in Transient Storage Device Attachments [Security], D. Rich, Apr., pp. 102-104. The Case for Flexible NIST Security Standards, F. Keblawi and D. Sullivan, June, pp. 19-26. Enterprise, Systems, and Software Engineering—The Need for Integration [Standards], P. Joannou, May, pp. 103-105. Fault-Tolerant CORBA: From Specification to Reality [Standards], P. Narasimhan, Jan., pp. 110-112. Software Process Improvement: The Competisoft Project, Oktaba H. et al., Oct., pp. 21-28. Standards, Agility, and Engineering, F. Coallier, Sept., pp. 100102. Standards Confusion and Harmonization [Standards], J.M. Voas and P.A. Laplante, July, pp. 94-96. Unsafe Standardization [Standards], M. Thomas, Nov., pp. 109-111. Stereo-vision systems 3D Vision: Developing an Embedded Stereo-Vision System [Embedded Computing], J.I. Woodfill et al., May, pp. 106108. Supercomputing The Green500 List: Encouraging Sustainable Supercomputing, W. C. Feng and K. Cameron, Dec., pp. 50-55. Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. Vforce: An Extensible Framework for Reconfigurable Supercomputing, N. Moore et al., Mar., pp. 39-49. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Swarm technologies Swarms and Swarm Intelligence [Software Technologies], M.G. Hinchey et al., Apr., pp. 111-113. System design and development A Communication Support System for Older People with Dementia, N. Alm et al., May, pp. 35-41. The Discipline of Embedded Systems Design, T.A. Henzinger and J. Sifakis, Oct., pp. 32-40. Holistic Sensing and Active Displays for Intelligent Driver Support Systems, M.M. Trivedi and S.Y. Cheng, May, pp. 60-68. How Business Goals Drive Architectural Design [Software Technologies], R.S. Sangwan and C.J. Neill, Aug., pp. 8587. Human-Centered Computing: Toward a Human Revolution [Guest Editors’ Introduction], A. Jaimes et al., May, pp. 3034. An Interactive Multimedia Diary for the Home, G.C. de Silva et al., May, pp. 52-59. T Tablet PCs Classroom Presenter: Enhancing Interactive Education with Digital Ink, R. Anderson et al., Sept., pp. 56-61. Facilitating Pedagogical Practices through a Large-Scale Tablet PC Deployment, J.G. Tront, Sept., pp. 62-68. Handwriting Recognition: Tablet PC Text Input, J.A. Pittman, Sept., pp. 49-54. Ink, Improvisation, and Interactive Engagement: Learning with Tablets, J. Roschelle et al., Sept., pp.42-48. Magic Paper: Sketch-Understanding Research, R. Davis, Sept., pp. 34-41. Tablet PC Technology: The Next Generation [Guest Editors’ Introduction], J. Prey and A. Weaver, Sept., pp. 32-33. Telecommunications industry Evolution of SOA Concepts in Telecommunications, T. Magedanz, et al., Nov., pp. 46-50. Transient storage devices Authentication in Transient Storage Device Attachments [Security], D. Rich, Apr., pp. 102-104. Trident compiler Trident: From High-Level Language to Hardware Circuitry, J.L. Tripp et al., Mar., pp. 28-37. Trust management Trust Management in Distributed Systems, H. Li and M. Singhal, Feb., pp. 45-53. U User-centered system design Human-Centered Computing: Toward a Human Revolution [Guest Editors’ Introduction], A. Jaimes et al., May, pp. 3034. User interfaces Enhancing the User Experience in Mobile Phones [Entertainment Computing], S.R. Subramanya and B.K. Yi, Dec., pp. 114-117. V Vector architecture Embracing and Extending 20th-Century Instruction Set Architectures, J. Gebis and D. Patterson, Apr., pp. 68-75. Vforce framework Vforce: An Extensible Framework for Reconfigurable Supercomputing, N. Moore et al., Mar., pp. 39-49. Virtual reality Generation 3D: Living in Virtual Worlds [Entertainment Computing], M. Macedonia, Oct., pp. 99-101. Immersidata Analysis: Four Case Studies, C. Shahabi et al., July, pp. 45-52. Powering Down the Computing Infrastructure [News], G. Lawton, Feb., pp. 16-19. Virtual Reality: How Much Immersion Is Enough?, D.A. Bowman and R.P. McMahan, July, pp. 36-43. Virtual Reality Program Eases Amputees’ Phantom Pain, L.D. Paulson, Jan., pp. 22-24. Volumetric displays 3D Display Using Passive Optical Scatterers, S.K. Nayar and V.N. Anand, July, pp. 54-63. VPNs A New Virtual Private Network for Today’s Mobile World, [Technology News], K. Heyman, Dec., pp. 17-19. A BEMaGS F W WS-QoS framework Adaptive QoS for Mobile Web Services through Cross-Layer Communication, M. Tian et al., Feb., pp. 59-63. Web 2.0 Building Web 2.0 [Web Technologies], K.-J. Lin, May, pp. 101-102. Enterprise Security for Web 2.0 [IT Systems Perspectives], M.A. Davidson and E. Yoran, Nov., pp. 117-119. Five Enablers for Mobile 2.0 [Invisible Computing], W.G. Griswold, Oct., pp. 96-98. Toward a Social Semantic Web [Web Technologies], A. Mikroyannidis, Nov., pp. 113-115. Web 2.0 Creates Security Challenges [News], G. Lawton, Oct., pp. 13-16. Web applications How-To Web Pages [Invisible Computing], C. Torrey and D.W. McDonald, Aug., pp. 96-97. Replacing Proprietary Software on the Desktop [IT Systems Perspectives], D. Hardaway, Mar., pp. 96-97. Social Scripting for the Web [Invisible Computing], T. Lau, June, pp. 96-98. Web services Adaptive QoS for Mobile Web Services through Cross-Layer Communication, M. Tian et al. , Feb., pp. 59-63. Component Contracts in Service-Oriented Architectures, F. Curbera, Nov., pp. 74-80. Evolution of SOA Concepts in Telecommunications, T. Magedanz, et al., Nov., pp. 46-50. The Fractal Nature of Web Services [Web Technologies], C. Bussler, Mar., pp. 93-95. Service Is in the Eyes of the Beholder, T. Margaria, Nov., pp. 33-37. Service-Oriented Computing: State of the Art and Research Challenges, M.P. Papazoglou et al., Nov., pp. 38-45. Toward Mobile Services: Three Approaches, J. Bosch, Nov., pp. 51-53. Toward the Realization of Policy-Oriented Enterprise Management, M. Kaiser, Nov., pp. 57-63. Web technologies Building Web 2.0 [Web Technologies], K.-J. Lin, May, pp. 101-102. Click Fraud [Web Technologies], B.J. Jansen, July, pp. 85-86. The Fractal Nature of Web Services [Web Technologies], C. Bussler, Mar., pp. 93-95. An Information Avalanche [Web Technologies], V.G. Cerf, Jan., pp. 104-105. Online Experiments: Lessons Learned, R. Kohavi and R. Longbotham, Sept., pp. 103-105. Toward a Social Semantic Web [Web Technologies], A. Mikroyannidis, Nov., pp. 113-115. Wi-Fi News Briefs, L.D. Paulson, July, pp. 17-19. Wi-Fi—The Nimble Musician in Your Laptop [How Things Work], D.G. Leeper, Apr., pp. 108-110. Wireless technology 4G Wireless Begins to Take Shape [News], S. Ortiz Jr., Nov., pp. 18-21. iMouse: An Integrated Mobile Surveillance and Wireless Sensor System, Y.-C. Tseng et al., June, pp. 60-66. Writing systems The Profession as a Culture Killer, [The Profession], N. Holmes, Sept., pp. 112, 110-111. AUTHOR INDEX A Agarwal, P.K., see S.R. Alam, Mar., pp. 66-73. Aggarwal, N., P. Ranganathan, N.P. Jouppi, and J.E. Smith, Isolation in Commodity Multicore Processors, June, pp. 49-59. Aiken, P., M.D Allen, B. Parker, and A. Mattia, Measuring Data Management Practice Maturity: A Community’s SelfAssessment, Apr., pp. 42-50. Aizawa, K., see G.C. de Silva, May, pp. 52-59. Akella, V., see J. Oliver, Dec., pp. 56-61. Alam, S.R., P.K. Agarwal, M.C. Smith, J.S. Vetter, and D. Caliga, Using FPGA Devices to Accelerate Biomolecular Simulations, Mar., pp. 66-73. Allen, M. David., see P. Aiken, Apr., pp. 42-50. 79 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX Alm, N., R. Dye, G. Gowans, J. Campbell, A. Astell, and M. Ellis, A Communication Support System for Older People with Dementia, May, pp. 35-41. Aloimonos, Y., see G. Guerra-Filho, May, pp. 42-51. Alquicira, C., see H. Oktaba, Oct., pp. 21-28. Amirtharajah, R., see J. Oliver, Dec., pp. 56-61. Anand, V.N., see S.K. Nayar, July, pp. 54-63. Anderson, R., R. Anderson, P. Davis, N. Linnell, C. Prince, V. Razmov, and F. Videon, Classroom Presenter: Enhancing Interactive Education with Digital Ink, Sept., pp. 56-61. Anderson, R., see R. Anderson, Sept., pp. 56-61. Astell, A., see N. Alm, May, pp. 35-41. B Bahar, R. Iris., D. Hammerstrom, J. Harlow, W.H. Joyner Jr., C. Lau, D. Marculescu, A. Orailoglu, and M. Pedram, Architectures for Silicon Nanoelectronics and Beyond, Jan., pp. 25-33. Bailey, J., see J. Spohrer, Jan., pp. 71-77. Baluja, S., see M. Kamvar, Aug., pp. 58-62. Bannon, L.J., see L. Ciolfi, July, pp. 64-71. Barroso , L. and U. Hölzle, The Case for Energy-Proportional Computing, Dec., pp. 33-37. Bay, J., see J. Sztipanovits, Mar., pp. 90-92. Berenbach, B., see J. Cleland-Huang, June, pp. 27-35. Berk, V. H., see G. Cybenko, Jan., pp. 62-70. Bertino, E., see J. Yong, Oct., pp. 89-91. Bhatia, L., see V. Vijairaghavan, Feb., pp. 54-58. Blum, N., see T. Magedanz, Nov., pp. 46-50. Bohner, S., An Era of Change-Tolerant Systems, June, pp. 100-102. Bosch, J., S. Friedrichs, S. Jung, J. Helbig, and A. Scherdin, Service Orientation in the Enterprise, Nov., pp. 51-56. Bowman, D.A., and R.P. McMahan, Virtual Reality: How Much Immersion Is Enough?, July, pp. 36-43. Brauner, D.F., see M.A. Casanova, Oct., pp. 102104. Breitman, K.K., see M.A. Casanova, Oct., pp. 102-104. Brereton, P., see D. Budgen, Apr., pp. 34-41. Brown, J., B. Shipman, and R. Vetter, SMS: The Short Message Service, Dec., pp. 106-110. Brown, T., see J.I. Woodfill, May, pp. 106-108. Buchta, J., see M. Petrenko, Nov., pp. 25-31. Buck, R., see J.I. Woodfill, May, pp. 106-108. Budgen, D., M. Rigby, P. Brereton, and M. Turner, A Data Integration Broker for Healthcare Systems, Apr., pp. 34-41. Buehler, M., see G. Dudek, Jan., pp. 46-53. Buell, D., T. El-Ghazawi, K. Gaj, and V. Kindratenko, Guest Editors’ Introduction: High-Performance Reconfigurable Computing, Mar., pp. 23-27. Burgstahler, S.E., and R.E. Ladner, Increasing the Participation of People with Disabilities in Computing Fields, May, pp. 94-97. Bussler, C., The Fractal Nature of Web Services, Mar., pp. 9395. C Caliga, D., see S.R. Alam, Mar., pp. 66-73. Cameron, K., see W. Feng , Dec., pp. 50-55. Campbell, J., see N. Alm, May, pp. 35-41. Cao, G., see J. Cao, Apr., pp. 60-66. Cao, J., Y. Zhang, G. Cao, and L. Xie, Data Consistency for Cooperative Caching in Mobile Environments, Apr., pp. 60-66. Carlin, A., and F. Gallegos, IT Audit: A Critical Business Process, July, pp. 87-89. Casanova, M.A., K.K. Breitman, D.F. Brauner, and A.L.. Marins, Database Conceptual Schema Matching, Oct., pp. 102-104. Cerf, V.G., An Information Avalanche, Jan., pp. 104-105. Chakrabarti, S., and M. Singhal, Password-Based Authentication: Preventing Dictionary Attacks, June, pp. 68-74. Chang, C.K. Computer Recognizes Expert Reviewers, Dec., pp. 64-65. My Vision for Computer, Jan., pp. 7-8. Chaudhury, S. Raj., see J. Roschelle, Sept., pp. 42-48. Cheng, K., see Y. Tseng, June, pp. 60-66. Cheng, S.Y., see M.M. Trivedi, May, pp. 60-68. Chong, F., see J. Oliver, Dec., pp. 56-61. 80 Ciolfi, L., M. Fernstrom, L.J. Bannon, P. Deshpande, P. Gallagher, C. McGettrick, N. Quinn, and S. Shirley, The Shannon Portal Installation: Interaction Design for Public Places, July, pp. 64-71. Clacy, B., and B. Jennings, Service Management: Driving the Future of IT, May, pp. 98-100. Clark, S., see J. Cleland-Hang, June, pp. 27-35. Cleland-Huang, J., B. Berenbach, S. Clark, R. Settimi, and E. Romanova, Best Practices for Automated Traceability, June, pp. 27-35. Coallier, F., Standards, Agility, and Engineering, Sept., pp. 100-102. Connelly, K., see K.A. Siek, Feb., pp. 89-92. Conti, A., see M.C. Herbordt, Mar., pp. 50-57. see N. Moore, Mar., pp. 39-49. Croson, D., see D. Schuff, Feb., pp. 31-36. Curbera, F., Component Contracts in Service-Oriented Architectures, Nov., pp. 74-80. Cybenko, G., and V.H. Berk, Process Query Systems, Jan., pp. 62-70. D D’Arcy, J., see D. Schuff, Feb., pp. 31-36. Davidson, M.A., and E. Yoran, Enterprise Security for Web 2.0, Nov., pp. 117-119. Davis, P., see R. Anderson, Sept., pp. 56-61. Davis, R., Magic Paper: Sketch-Understanding Research, Sept., pp. 34-41. de Silva, G.C., T. Yamasaki, and K. Aizawa, An Interactive Multimedia Diary for the Home, May, pp. 52-59. Deelman, E., see Y. Gil, Dec., pp. 24-32. Deshpande, P., see L. Ciolfi, July, pp. 64-71. Di Natale, M., see A. Sangiovanni-Vincentelli, Oct., pp. 4251. DiGiano, C., see J. Roschelle, Sept., pp. 42-48. Dimitriadis, Y., see J. Roschelle, Sept., pp. 42-48. DiSabello, D., see M.C. Herbordt, Mar., pp. 50-57. Dudek, G., P. Giguere, C. Prahacs, S. Saunderson, J. Sattar, L. Torres-Mendez, M. Jenkin, A. German, A. Hogue, A. Ripsman, J. Zacher, E. Milios, H. Liu, P. Zhang, M. Buehler, and C. Georgiades, AQUA: An Amphibious Autonomous Robot, Jan., pp. 46-53. Dustdar, S., see M.P. Papazoglou, Nov., pp. 38-45. Dutkowski, S., see T. Magedanz, Nov., pp. 46-50. Dye, R., see N. Alm, May, pp. 35-41. E El-Ghazawi, T., see D. Buell, Mar., pp. 23-27. Ellis, M., see N. Alm, May, pp. 35-41. Ellisman, M., see Y. Gil, Dec., pp. 24-32. Epstein, J., Electronic Voting, Aug., pp. 92-95. Erb, D., see M. Gschwind, June, pp. 37-47. F Fahringer, T., see Y. Gil, Dec., pp. 24-32. Fang, F., see M. Parameswaran, Jan., pp. 40-44. Feng , W., and K. Cameron, The Green500 List: Encouraging Sustainable Supercomputing, Dec., pp. 50-55. Fernstrom, M., see L. Ciolfi, July, pp. 64-71. Fiadeiro, J.L., Designing for Software’s Social Complexity, Jan., pp. 34-39. Flinn, J., see D. Peek, Feb., pp. 93-95. Fox, G., see Y. Gil, Dec., pp. 24-32. Friedrichs, S., see J. Bosch, Nov., pp. 51-56. G Gaj, K., see D. Buell, Mar., pp. 23-27. Galgali, P., see V. Vijairaghavan, Feb., pp. 54-58. Gallagher, P., see L. Ciolfi, July, pp. 64-71. Gallegos, F., see A. Carlin, July, pp. 87-89. Gannon, D., see Y. Gil, Dec., pp. 24-32. Garcia, F., see H. Oktaba, Oct., pp. 21-28. Gatica-Perez, D., see A. Jaimes, May, pp. 30-34. Gaubatz, G., see J. Kaps, Feb., pp. 38-44. Gebis, J., and D. Patterson, Embracing and Extending 20thCentury Instruction Set Architectures, Apr., pp. 68-75. Geer, D. [News] Improving Data Accessibility with File Area Networks, Nov., pp. 14-17. For Programmers, Multicore Chips Mean Multiple Challenges, Sept., pp. 17-19. Georgiades, C., see G. Dudek, Jan., pp. 46-53. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page German, A., see G. Dudek, Jan., pp. 46-53. Geyer, R., see J. Oliver, Dec., pp. 56-61. Giguere, P., see G. Dudek, Jan., pp. 46-53. Gil, Y., E. Deelman, M. Ellisman, T. Fahringer, G. Fox, D. Gannon, C. Goble, M. Livny, L. Moreau, and J. Myers, Examining the Challenges of Scientific Workflows, Dec., pp. 24-32. Goble, C., see Y. Gil, Dec., pp. 24-32. Gokhale, M.B., see J.L. Tripp, Mar., pp. 28-37. Golston, J., see D. Talla, Oct., pp. 53-61. Gordon, G., see J. Iselin. Woodfill, May, pp. 106-108. Gowans, G., see N. Alm, May, pp. 35-41. Gramm, A., see M. Tian, Feb., pp. 59-63. Grier, D.A., [Column] Annie and the Boys, Aug., pp. 6-9. The Best Deal in Town, Apr., pp. 8-11. The Boundaries of Time, July, pp. 5-7. The Camino Real, June, pp. 6-8. Controlling the Conversation, Sept., pp. 7-9. Counting Beans, Nov., pp. 8-10. Dirty Electricity, Feb., pp. 6-8. E-Mailing from Armenia, Oct., pp. 8-10. A Force of Nature, Dec., pp. 8-10. Outposts, Mar., pp. 8-10. The Wave of the Future, Jan., pp. 12-14. Working Class Hero, May, pp. 8-10. Griswold, W. G., Five Enablers for Mobile 2.0, Oct., pp. 9698. Grottke, M., and K. S. Trivedi, Fighting Bugs: Remove, Retry, Replicate, and Rejuvenate, Feb., pp. 107-109. Gruhl, D., see J. Spohrer, Jan., pp. 71-77. Gschwind, M., D. Erb, S. Manning, and M. Nutter, An Open Source Environment for Cell Broadband Engine System Software, June, pp. 37-47. Gu, Y., see M.C. Herbordt, Mar., pp. 50-57. Guerra-Filho, G., and Y. Aloimonos, A Language for Human Action, May, pp. 42-51. H Hammerstrom, D., see R. Iris. Bahar, Jan., pp. 25-33. Hardaway, D., Replacing Proprietary Software on the Desktop, Mar., pp. 96-97. Harkiolakis, N., Incorporating a Variable-Expertise-Level System in IT Course Modules, Apr., pp. 116, 114-115. Harlow, J., see R. Iris. Bahar, Jan., pp. 25-33. Hars, L., Discryption: Internal Hard-Disk Encryption for Secure Storage, June, pp. 103-105. Hatton, L. The Chimera of Software Quality, Aug., pp. 104, 102-103. Empirical Test Observations in Client-Server Systems, May, pp. 24-29. How Accurately Do Engineers Predict Software Maintenance Tasks?, Feb., pp. 64-69. Helbig, J., see J. Bosch, Nov., pp. 51-56. Henzinger, T.A., and J. Sifakis, The Discipline of Embedded Systems Design, Oct., pp. 32-40. Herbordt, M.C., T. VanCourt, Y. Gu, B. Sukhwani, A. Conti, J. Model, and D. DiSabello, Achieving High Performance with FPGA-Based Computing, Mar., pp. 50-57. Heyman, K. [News] New Attack Tricks Antivirus Software, May, pp. 18-20. A New Virtual Private Network for Today’s Mobile World, Dec., pp. 17-19. Hinchey, M.G., R. Sterritt, and C. Rouff, Swarms and Swarm Intelligence, Apr., pp. 111-113. Hogue, A., see G. Dudek, Jan., pp. 46-53. Hole, K.J., see A.N. Klingsheim, Feb., pp. 24-30. Holmes, N. [Column] Binary Arithmetic, June, pp. 90-93. The Computing Profession and Higher Education, Jan., pp. 116, 114-115. Consciousness and Computers, July, pp. 100, 98-99. Digital Technology and the Skills Shortage, Mar., pp. 100, 98-99. The Profession as a Culture Killer, Sept., pp. 112, 110-111. Holzmann, G., Conquering Complexity, Dec., pp. 111-113. Hopkins, L., see K.A. Siek, Feb., pp. 89-92. Horvitz, E., see J. Krumm, Apr., pp. 105-107. Hsieh, Y., see Y. Tseng, June, pp. 60-66. Huang, T.S., see A. Jaimes, May, pp. 30-34. Humphreys, G., see D. Luebke, Feb., pp. 96-100. Hölzle, U., see L. Barroso , Dec., pp. 33-37. A BEMaGS F J Jaimes, A., D. Gatica-Perez, N. Sebe, and T.S. Huang, Guest Editors’ Introduction: Human-Centered Computing— Toward a Human Revolution, May, pp. 30-34. Jansen, B.J., Click Fraud, July, pp. 85-86. Jansen, B.J., and A. Spink, Sponsored Search: Is Money a Motivator for Providing Relevant Results?, Aug., pp. 52-57. Jenkin, M., see G. Dudek, Jan., pp. 46-53. Jennings, B., see B. Clacy, May, pp. 98-100. Joachims, T., and F. Radlinski, Search Engines that Learn from Implicit Feedback, Aug., pp. 34-40. Joannou, P., Enterprise, Systems, and Software Engineering— The Need for Integration, May, pp. 103-105. Jouppi, N.P., see N. Aggarwal, June, pp. 49-59. Joyner Jr., W.H., see R. Iris. Bahar, Jan., pp. 25-33. Jung, S., see J. Bosch, Nov., pp. 51-56. Jurasek, D., see J.I. Woodfill, May, pp. 106-108. K Kaiser, M., Toward the Realization of Policy-Oriented Enterprise Management, Nov., pp. 57-63. Kamvar, M., and S. Baluja, Deciphering Trends in Mobile Search, Aug., pp. 58-62. Kaps, J., G. Gaubatz, and B. Sunar, Cryptography on a Speck of Dust, Feb., pp. 38-44. Keblawi, F., and D. Sullivan, The Case for Flexible NIST Security Standards, June, pp. 19-26. Kindratenko, V., see D. Buell, Mar., pp. 23-27. King, B.A., and L.D. Paulson, Motion Capture Moves into New Realms, Sept., pp. 13-16. King, L.S., see N. Moore, Mar., pp. 39-49. Klingsheim, A.N., V. Moen, and K.J. Hole, Challenges in Securing Networked J2ME Applications, Feb., pp. 24-30. Knight, J., The Glass Cockpit, Oct., pp. 92-95. Kohavi, R., and R. Longbotham, Online Experiments: Lessons Learned, Sept., pp. 103-105. Kozyrakis, C., see S. Rivoire, Dec., pp. 39-48. Krumm, J., and E. Horvitz, Predestination: Where Do You Want to Go Today?, Apr., pp. 105-107. L Ladner, R.E., see S.E. Burgstahler, May, pp. 94-97. Laird, C., Taking a Hard-Line Approach to Encryption, Mar., pp. 13-15. Laplante, P.A., see J.M. Voas, July, pp. 94-96. Lau, C., see R. Iris. Bahar, Jan., pp. 25-33. Lau, T., Social Scripting for the Web, June, pp. 9698. Lawton, G. [News] The Next Big Thing in Chipmaking, Apr., pp. 18-20. Powering Down the Computing Infrastructure, Feb., pp. 16-19. Stronger Domain Name System Thwarts Root-Server Attacks, May, pp. 14-17. These Are Not Your Father’s Widgets, July, pp. 10-13. Web 2.0 Creates Security Challenges, Oct., pp. 13-16. Leavitt, N. [News] The Changing World of Outsourcing, Dec., pp. 13-16. For Wireless USB, the Future Starts Now, July, pp. 14-16. Vendors Fight Spam’s Sudden Rise, Mar., pp. 16-19. Leeper, D. G., Wi-Fi—The Nimble Musician in Your Laptop, Apr., pp. 108-110. Leeser, M., see N. Moore, Mar., pp. 39-49. Leymann, F., see M. P. Papazoglou, Nov., pp. 38-45. Li, H., and M. Singhal, Trust Management in Distributed Systems, Feb., pp. 45-53. Lin, K., Building Web 2.0, May, pp. 101-102. Linnell, N., see R. Anderson, Sept., pp. 56-61. Liu, H., see G. Dudek, Jan., pp. 46-53. Liu, J., see S. Nath, July, pp. 90-93. Liu, J., Computing as an Evolving Discipline: 10 Observations, May, pp. 112, 110-111. Livny, M., see Y. Gil, Dec., pp. 24-32. Loka, R.R., Software Development: What Is the Problem?, Feb., pp. 112, 110-111. Longbotham, R., see R. Kohavi, Sept., pp. 103-105. Luebke, D., and G. Humphreys, How GPUs Work, Feb., pp. 96-100. M Macedonia, M. [Column] The Future Arrives ... Finally, Feb., pp. 101-103. 81 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX Generation 3D: Living in Virtual Worlds, Oct., pp. 99-101. iPhones Target the Tech Elite, June, pp. 94-95. Magedanz, T., N. Blum, and S. Dutkowski, Evolution of SOA Concepts in Telecommunications, Nov., pp. 46-50. Maglio, P.P., see J. Spohrer, Jan., pp. 71-77. Manning, S., see M. Gschwind, June, pp. 37-47. Marculescu, D., see R. Iris. Bahar, Jan., pp. 25-33. Margaria, T., Service Is in the Eyes of the Beholder, Nov., pp. 33-37. Marins, A.L.A., see M.A. Casanova, Oct., pp. 102-104. Marsh, T., see C. Shahabi, July, pp. 45-52. Mattia, A., see P. Aiken, Apr., pp. 42-50. Mayer, A., H. Siebert, and K.D. McDonald-Maier, Boosting Debugging Support for Complex Systems on Chip, Apr., pp. 76-81. McDonald, D.W., see C. Torrey, Aug., pp. 96-97. McDonald-Maier, K.D., see A. Mayer, Apr., pp. 76-81. McGettrick, C., see L. Ciolfi, July, pp. 64-71. McLaughlin, M., see C. Shahabi, July, pp. 45-52. McMahan, R.P., see D. A. Bowman, July, pp. 36-43. Menzel, S., see K.A. Siek, Feb., pp. 89-92. Menzies, T., D. Owen and J. Richardson, The Strangest Thing About Software, Jan., pp. 54-60. Messerges, T., see M. Tripunitara, Feb., pp. 104-106. Meza , J., see S. Rivoire, Dec., pp. 39-48. Mikroyannidis, A., Toward a Social Semantic Web, Nov., pp. 113-115. Milios, E., see G. Dudek, Jan., pp. 46-53. Model, J., see M.C. Herbordt, Mar., pp. 50-57. Moen, V., see A.N. Klingsheim, Feb., pp. 24-30. Moore, N., A. Conti, M. Leeser, and L.S. King, Vforce: An Extensible Framework for Reconfigurable Supercomputing, Mar., pp. 39-49. Moreau, L., see Y. Gil, Dec., pp. 24-32. Morris, G.R., and V.K. Prasanna, Sparse Matrix Computations on Reconfigurable Hardware, Mar., pp. 58-64. Mun, M., see C. Shahabi, July, pp. 45-52. Myers, J., see Y. Gil, Dec., pp. 24-32. N Narasimhan, P., Fault-Tolerant CORBA: From Specification to Reality, Jan., pp. 110-112. Narayan, P., see B. Steffen, Nov., pp. 64-73. Nath, S., J. Liu and F. Zhao, SensorMap for WideArea Sensor Webs, July, pp. 90-93. Nayar, S.K., and V.N. Anand, 3D Display Using Passive Optical Scatterers, July, pp. 54-63. Neill, C.J., see R.S. Sangwan, Aug., pp. 85-87. Nutter, M., see M. Gschwind, June, pp. 37-47. O Oktaba, H., F. Garcia, M. Piattini, F. Ruiz, F.J. Pino, and C. Alquicira, Software Process Improvement: The Competisoft Project, Oct., pp. 21-28. Oliver, J., R. Amirtharajah, V. Akella, R. Geyer and F. Chong, Life Cycle Aware Computing: Reusing Silicon Technology, Dec., pp. 56-61. Orailoglu, A., see R. Iris. Bahar, Jan., pp. 25-33. Ortiz Jr., S. [News] 4G Wireless Begins to Take Shape, Nov., pp. 18-21. Brain-Computer Interfaces: Where Human and Machine Meet, Jan., pp. 17-21. Getting on Board the Enterprise Service Bus, Apr., pp. 15-17. Protecting Networks by Controlling Access, Aug., pp. 16-19. Searching the Visual Web, June, pp. 12-14. Owen, D., see T. Menzies, Jan., pp. 54-60. P Papazoglou, M.P., P. Traverso, S. Dustdar, and F. Leymann, Service-Oriented Computing: State of the Art and Research Challenges, Nov., pp. 38-45. Parameswaran, M., X. Zhao, A.B. Whinston and F. Fang, Reengineering the Internet for Better Security, Jan., pp. 4044. Parker, B., see P. Aiken, Apr., pp. 42-50. Patterson, D., see J. Gebis, Apr., pp. 68-75. Patton, C., see J. Roschelle, Sept., pp. 42-48. Paul, J. M., see S. M. Pieper, Sept., pp. 23-30. Paulson, L.D. [News Briefs] Company Develops Handheld with Flexible Screen, Apr., p. 22. 82 Company Says Diagonal Wiring Makes Chips Faster, Sept., p. 20. Developing Tomorrow’s Beach Today, Where High Tide Meets High Tech, Oct., p. 18. Femtocells Promise Faster Mobile Networks, Oct., p. 18. Google Surveys Web for Malware, July, p. 18 High-Tech Mirror Helps Shoppers Reflect on Their Purchases Aug., p. 21. Hitachi Researchers Develop Power-Sized RFID Chips, May, p. 23. HP Releases Computer on a Sticker, Feb., p. 21 IBM Adds “Nothing” to Chips, Improves Performance, July, p. 17. IBM Demonstrates Fast Memory-Chip Technology, Apr., p. 22. IEEE Works on Energy-Efficient Ethernet, May, pp. 21. Important Wireless-Spectrum Auction Generates Controversy Nov. p. 22. Intel Adds Distance to Wi-Fi, July, p. 19. Interface Makes a Computer Desktop Look like a Real Desktop, Dec. p. 21. Light Pulses Could Improve Optical Communications, Oct., p. 17. Making Web Science an Academic Field, Feb., p. 22. Nanodevice Increases Optical Networks’ Bandwidth, Apr., p. 21. New Approach Will Speed Up USB, Dec., p. 21. New Attack Works Hards to Avoid Defenses, Aug., p. 20. New Contender in Networked Storage, Feb., p. 20. New Eye-Tracking Technology Could Make Billboards More Effective, July, p. 18. New Software Visually Displays Musical Structure; May, p. 22. New Technique Yields Faster Multicore Chips, Oct., p. 19. New Technology Creates Lively E-Wallpaper, Nov., p. 23. New Technology Prevents Click Fraud Mar., p. 21. New Technology Transmits Data via Visible Light, Sept., p. 21. Professor Builds Desktop High-Performance Computer, Nov., p. 22. Project Improves Mesh Networks, Jan., p. 23. Researchers Develop “Ballistic Computing” Transistor, Jan., p. 22. Researchers Develop Efficient Digital-Camera Design Mar., p. 20. Researchers Herd Computers to Fight Spyware, Sept., p. 22. Robot Adapts to Changing Conditions, Dec. p. 20. Scientists Study Ancient Computing Device, Feb., p. 21. Scrabble Program Wins by Inference, Mar., p. 21. Silicon Clock Promises Improved Computer Technology, Mar., p. 22. Standardization Comes to Virtualization, Dec., p. 22. Startup Uses Sensors to Save Houseplants, Apr., p. 23 Spherical System Captures Images from All Directions, Aug., p. 20. Sun Makes Java Open Source, Jan., p. 24. Technique Creates High-Performance Storage Technology, Aug., p. 21. Tracking Troubled Turtles with Wireless Technology, Sept., p. 21. Virtual Reality Program Eases Amputees’ Phantom Pain, Jan., p. 23. Will Thin Finally Be In?, Nov., p. 24. Pedram, M., see R. Iris. Bahar, Jan., pp. 25-33. Peek, D., and J. Flinn, Consumer Electronics Meets Distributed Storage, Feb., pp. 93-95. Peterson, K.D., see J.L. Tripp, Mar., pp. 28-37. Petrenko, M., D. Poshyvanyk, V. Rajlich, and J. Buchta, Teaching Software Evolution in Open Source, Nov., pp. 25-31. Piattini, M., see H. Oktaba, Oct., pp. 21-28. Pieper, S.M., J.M. Paul, and M.J. Schulte, A New Era of Performance Evaluation, Sept., pp. 23-30. Pino, F.J., see H. Oktaba, Oct., pp. 21-28. Pittman, J.A., Handwriting Recognition: Tablet PC Text Input, Sept., pp. 49-54. Poshyvanyk, D., see M. Petrenko, Nov., pp. 25-31. Prahacs, C., see G. Dudek, Jan., pp. 46-53. Prasanna, V.K., see G.R. Morris, Mar., pp. 58-64. Prey, J., and A. Weaver, Guest Editors’ Introduction: Tablet PC Technology—The Next Generation, Sept., pp. 32-33. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Price, S.M., Supporting Resource-Constrained Collaboration Environments, June, pp. 108, 106-107. Prince, C., see R. Anderson, Sept., pp. 56-61. Q Quinn, N., see L. Ciolfi, July, pp. 64-71. R Radlinski, F., see T. Joachims, Aug., pp. 34-40. Rajlich, V., see M. Petrenko, Nov., pp. 25-31. Ramacher, U., Software-Defined Radio Prospects for Multistandard Mobile Phones, Oct., pp. 62-69. Ramakrishnan, N., From the Area Editor: Search—The New Incarnations, Aug., pp. 31-32. Ramakrishnan, R., and A. Tomkins, Toward a PeopleWeb, Aug., pp. 63-72. Ranganathan, P., see S. Rivoire, Dec., pp. 39-48. see N. Aggarwal, June, pp. 49-59. Razmov, V., see R. Anderson, Sept., pp. 56-61. Rich, D., Authentication in Transient Storage Device Attachments, Apr., pp. 102-104. Richardson, J., see T. Menzies, Jan., pp. 54-60. Riehle, D., The Economic Motivation of Open Source Software: Stakeholder Perspectives, Apr., pp. 25-32. Rigby, M., see D. Budgen, Apr., pp. 34-41. Ripsman, A., see G. Dudek, Jan., pp. 46-53. Ritter, H., see M. Tian, Feb., pp. 59-63. Rivoire, S., M. Shah, P. Ranganathan, C. Kozyrakis, and J. Meza , Models and Metrics to Enable Energy-Efficiency Optimizations, Dec., pp. 39-48. Rizzo, A.A., see C. Shahabi, July, pp. 45-52. Rohrbough, L., see J. Sztipanovits, Mar., pp. 90-92. Romanova, E., see J. Cleland-Huang, June, pp. 27-35. Roschelle, J., D. Tatar, S. Raj. Chaudhury, Y. Dimitriadis, C. Patton, and C. DiGiano, Ink, Improvisation, and Interactive Engagement: Learning with Tablets, Sept., pp. 42-48. Ross, R., Managing Enterprise Security Risk with NIST Standards, Aug., pp. 88-91. Rouff, C., see M.G. Hinchey, Apr., pp. 111-113. Ruiz, F., see H. Oktaba, Oct., pp. 21-28. S Sangiovanni-Vincentelli, A., and M. Di Natale, Embedded System Design for Automotive Applications, Oct., pp. 42-51. Sangwan, R.S., and C.J. Neill, How Business Goals Drive Architectural Design, Aug., pp. 85-87. Santini, S., Making Computers Do More with Less, Dec., pp. 124, 122-123. Sastry, S., see J. Sztipanovits, Mar., pp. 90-92. Sattar, J., see G. Dudek, Jan., pp. 46-53. Saunderson, S., see G. Dudek, Jan., pp. 46-53. Scherdin, A., see J. Bosch, Nov., pp. 51-56. Schiller, J.H., see M. Tian, Feb., pp. 59-63. Schmidt, D.C., see J. Sztipanovits, Mar., pp. 90-92. Schuff, D., O. Turetken, J. D’Arcy, and D. Croson, Managing E-Mail Overload: Solutions and Future Challenges, Feb., pp. 31-36. Schulte, M.J., see S.M. Pieper, Sept., pp. 23-30. Sebe, N., see A. Jaimes, May, pp. 30-34. Settimi, R., see J. Cleland-Huang, June, pp. 27-35. Shah, A., see V. Vijairaghavan, Feb., pp. 54-58. Shah, D., see V. Vijairaghavan, Feb., pp. 54-58. Shah, M., see S. Rivoire, Dec., pp. 39-48. Shah, N., see V. Vijairaghavan, Feb., pp. 54-58. Shahabi, C., T. Marsh, K. Yang, H. Yoon, A.A. Rizzo, M. McLaughlin, and M. Mun, Immersidata Analysis: Four Case Studies, July, pp. 45-52. Sharkey, N., Automated Killers and the Computing Profession, Nov., pp. 124, 122-123. Shipman, B., see J. Brown, Dec., pp. 106-110. Shirley, S., see L. Ciolfi, July, pp. 64-71. Siebert, H., see A. Mayer, Apr., pp. 76-81. Siek, K. A., K. Connelly, S. Menzel, and L. Hopkins, Propagating Diversity through Active Dissemination, Feb., pp. 89-92. Sifakis, J., see T.A. Henzinger, Oct., pp. 32-40. Singhal, M., see H. Li, Feb., pp. 45-53. Singhal, M., see S. Chakrabarti, June, pp. 68-74. Smith, J.E., see N. Aggarwal, June, pp. 49-59. Smith, M.C., see S.R. Alam, Mar., pp. 66-73. Smyth, B., A Community-Based Approach to Personalizing Web Search, Aug., pp. 42-50. A BEMaGS F Soukup, J., and M. Soukup, The Inevitable Cycle: Graphical Tools and Programming Paradigms, Aug., pp. 24-30. Soukup, M., see J. Soukup, Aug., pp. 24-30. Sousa, L., see S. Yamagiwa, May, pp. 70-77. Spink, A., see B.J. Jansen, Aug., pp. 52-57. Spohrer, J., P.P. Maglio, J. Bailey, and D. Gruhl, Steps Toward a Science of Service Systems, Jan., pp. 71-77. Srinivasan, V., see V. Vijairaghavan, Feb., pp. 54-58. Steffen, B., and P. Narayan, Full Life-Cycle Support for Endto-End Processes, Nov., pp. 64-73. Sterritt, R., see M.G. Hinchey, Apr., pp. 111-113. Stone, A., Natural-Language Processing for Intrusion Detection, Dec., pp. 103-105. Subramanya , S., and B. Yi, Enhancing the User Experience in Mobile Phones, Dec., pp. 114-117. Sukhwani, B., see M.C. Herbordt, Mar., pp. 50-57. Sullivan, D., see F. Keblawi, June, pp. 19 -26. Sunar, B., see J. Kaps, Feb., pp. 38-44. Swartout, W., see M. van Lent, Aug., pp. 98-100. Sztipanovits, J., J. Bay, L. Rohrbough, S. Sastry, D.C. Schmidt, N. Whitaker, D. Wilson, and D. Winter, Escher: A New Technology Transitioning Model, Mar., pp. 90-92. T Talla, D., and J. Golston, Using DaVinci Technology for Digital Video Devices, Oct., pp. 53-61. Tatar, D., see J. Roschelle, Sept., pp. 42-48. Thomas, M., Unsafe Standardization, Nov., pp. 109-111. Tian, M., A. Gramm, H. Ritter, J.H. Schiller, and T. Voigt, Adaptive QoS for Mobile Web Services through Cross-Layer Communication, Feb., pp. 59-63. Tomkins, A., see R. Ramakrishnan, Aug., pp. 6372. Torres-Mendez, L., see G. Dudek, Jan., pp. 46-53. Torrey, C., and D.W. McDonald, How-To Web Pages, Aug., pp. 96-97. Traverso, P., see M.P. Papazoglou, Nov., pp. 38-45. Treleaven, P., and J. Wells, 3D Body Scanning and Healthcare Applications, July, pp. 28-34. Tripp, J.L., M.B. Gokhale, and K.D. Peterson, Trident: From High-Level Language to Hardware Circuitry, Mar., pp. 28-37. Tripunitara, M., and T. Messerges, Resolving the Micropayment Problem, Feb., pp. 104-106. Trivedi, K.S., see M. Grottke, Feb., pp. 107-109. Trivedi, M.M., and S.Y. Cheng, Holistic Sensing and Active Displays for Intelligent Driver Support Systems, May, pp. 60-68. Tront, J.G., Facilitating Pedagogical Practices through a Large-Scale Tablet PC Deployment, Sept., pp. 62-68. Tseng, Y., Y. Wang, K. Cheng, and Y. Hsieh, iMouse: An Integrated Mobile Surveillance and Wireless Sensor System, June, pp. 60-66. Turetken, O., see D. Schuff, Feb., pp. 31-36. Turner, M., see D. Budgen, Apr., pp. 34-41. V Vahid, F., It’s Time to Stop Calling Circuits Hardware, Sept., pp. 106-108. VanCourt, T., see M.C. Herbordt, Mar., pp. 50-57. van Genuchten, M., The Impact of Software Growth on the Electronics Industry, Jan., pp. 106-108. van Genuchten, M., and D. Vogel, Getting Real in the Classroom, Oct., pp. 108, 106-107. Van Lent, M., Game Smarts, Apr., pp. 99-101. Van Lent, M., and W. Swartout, Games: Once More, with Feeling, Aug., pp. 98-100. Vasconcelos, N., From Pixels to Semantic Spaces: Advances in Content-Based Image Retrieval, July, pp. 20-26. Vaughan Nichols, S.J., New Interfaces at the Touch of a Fingertip, Aug., pp. 12-15. Vetter, J.S., see S. R. Alam, Mar., pp. 66-73. Vetter, R., see J. Brown, Dec., pp. 106-110. Videon, F., see R. Anderson, Sept., pp. 56-61. Vijairaghavan, V., D. Shah, P. Galgali, A. Shah, N. Shah, V. Srinivasan, and L. Bhatia, Marking Technique to Isolate Boundary Router and Attacker, Feb., pp. 54-58. Vincent, L., Taking Online Maps Down to Street Level, Dec., pp. 118-120. Voas, J.M., and P.A. Laplante, Standards Confusion and Harmonization, July, pp. 94-96. 83 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ANNUAL INDEX Vogel, D., see M. van Genuchten, Oct., pp. 108, 106-107. Voigt, T., see M. Tian, Feb., pp. 59-63. W Wang, Y., see Y. Tseng, June, pp. 60-66. Ward, B. [CS Connection] Bylaws Changes, Jan., pp. 89-93. CHC61 Sites Highlight ‘Unsung Heroes’ in 2007, Mar., pp. 77-79. Computer Science Enrollments Drop in 2006, June, p. 85. Computer Society Announces 2007 Programs, Jan., pp. 8487. Computer Society Announces Larson, UPE, and OCA Student Winners, Apr., p. 91. Computer Society Magazines: 2008 Preview, Nov., pp. 8184. Computer Society Recognizes Outstanding Professionals, May, p. 88. Computer Society Summer and Fall Conferences, May, pp. 85-87. Edward Seidel Honored with Sidney Fernbach Award, Apr., p. 89. Hosaka and Spielberg Named Winners of 2006 Computer Pioneer Award, Feb. p. 73. IEEE Computer Society Appoints Editors in Chief, Aug., pp. 77-78. IEEE Computer Society e-Learning Campus Adds Value to Membership, Oct., pp. 73-74. IEEE Computer Society Launches New Certification Program, Dec., pp. 64-65 IEEE Computer Society Offers College Scholarships, Aug., p. 79. IEEE Names 2007 Fellows, Feb., pp. 74-75. IEEE President-Elect Candidates Address Computer Society Concerns, Sept., pp. 77-83. James Pomerene Garners Joint IEEE/ACM Award, Apr., p. 90. Mateo Valero Receives Joint IEEE/ACM Award, Aug., p. 77. Oregon Teen Wins Computer Society Prize at Intel Science Fair, July, pp. 77-78. Society Board Amends Bylaws, July, pp. 79-80. Society Honors Wilkes and Parnas with 60th Anniversary Award, Nov., p. 85. Society Introduces New Technical Task Forces, June, p. 84. Society Publications Seek Editors in Chief for 2008-2009 Terms, Jan., p. 88. Software Engineering Volume Translated into Russian, Oct., p. 75. Tadashi Watanabe Receives 2006 Seymour Cray Award, Apr., p. 89. UCSD’s Smarr Receives Kanai Award, Oct. p. 73. Watson, H.J., and B.H. Wixom, The Current State of Business Intelligence, Sept., pp. 96-99. Weaver, A., see J. Prey, Sept., pp. 32-33. Wells, J., see P. Treleaven, July, pp. 28-34. Whinston, A B., see M. Parameswaran, Jan., pp. 40-44. Whitaker, N., see J. Sztipanovits, Mar., pp. 90-92. Williams, M.R. An Interesting Year, Dec., pp. 6-7. A Year of Decision, Jan., pp. 9-11. Wilson, D., see J. Sztipanovits, Mar., pp. 90-92. Winter, D., see J. Sztipanovits, Mar., pp. 90-92. Wixom, B. H., see H. J. Watson, Sept., pp. 96-99. Wolf, W. The Good News and the Bad News, Nov., pp. 104-105. Guest Editor’s Introduction: The Embedded Systems Landscape, Oct., pp. 29-31. Woodfill, J., R. Buck, D. Jurasek, G. Gordon, and T. Brown, 3D Vision: Developing an Embedded Stereo-Vision System, May, pp. 106-108. X Xie, L., see J. Cao, Apr., pp. 60-66. Y Yamagiwa, S., and L. Sousa, Caravela: A Novel Stream-Based Distributed Computing Environment, May, pp. 70-77. Yamasaki, T., see G. C. de Silva, May, pp. 52-59. Yang, K., see C. Shahabi, July, pp. 45-52. Yi, B., see S. Subramanya , Dec., pp. 114-117. Yong, J., and E. Bertino, Replacing Lost or Stolen E-Passports, Oct., pp. 89-91. Yoon, H., see C. Shahabi, July, pp. 45-52. Yoran, E., see M. Ann. Davidson, Nov., pp. 117-119. Z Zacher, J., see G. Dudek, Jan., pp. 46-53. Zhang, N., and W. Zhao, Privacy-Preserving Data Mining Systems, Apr., pp. 52-58. Zhang, P., see G. Dudek, Jan., pp. 46-53. Zhang, Y., see J. Cao, Apr., pp. 60-66. Zhao, F., see S. Nath, July, pp. 90-93. Zhao, W., see N. Zhang, Apr., pp. 52-58. Zhao, X., see M. Parameswaran, Jan., pp. 40-44. Join the IEEE Computer Society online at www.computer.org/join/ Complete the online application and get • immediate online access to Computer • a free e-mail alias — you@computer.org _________________ • free access to 100 online books on technology topics • free access to more than 100 distance learning course titles • access to the IEEE Computer Society Digital Library for only $118 Read about all the benefits of joining the Society at www.computer.org/join/benefits.htm 84 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ADVERTISER / PRODUCT INDEX ADVERTISER / PRODUCT INDEX D EC EM B ER Advertisers Page Number Alfaisal University AVSS 2008 87 117 Chinese University of Hong Kong CTS 2008 88 120 IEEE Computer Society Digital Library IEEE Computer Society Membership The George Washington University ieee.tv ____ 5 68-70 Lincoln Laboratory 97 Michigan State University 87 Microsoft 16 National University of Singapore 94 Philips 86 Seapine Software, Inc. Cover 4 SCC 2008 Cover 3 Temple University 87, 97 University of Bridgeport 92 University of California, Riverside 94 University of Massachusetts Boston 91 University of Michigan, Ann Arbor 90 University of Southern California 95 University of Toronto 89 University of Washington Tacoma 93 West Virginia University 91 Classified Advertising Advertising Sales Representatives Mid Atlantic (product/recruitment) Dawn Becker Phone: +1 732 772 0160 Fax: +1 732 772 0164 Email: ____________ db.ieeemedia@ieee.org Midwest/Southwest (recruitment) Darcy Giovingo Phone: +1 847 498 4520 Fax: +1 847 498 5911 ____________ Email: dg.ieeemedia@ieee.org New England (product) Jody Estabrook Phone: +1 978 244 0192 Fax: +1 978 244 0103 Email: ____________ je.ieeemedia@ieee.org Southwest (product) Steve Loerch Phone: +1 847 498 4520 Fax: +1 847 498 5911 Email: steve@didierandbroderick.com ________________ 92 Cover 2 86-101 200 7 New England (recruitment) John Restchack Phone: +1 212 419 7578 Fax: +1 212 419 7589 Email: ___________ j.restchack@ieee.org Northwest (product) Lori Kehoe Phone: +1 650-458-3051 Fax: +1 650 458 3052 Email: _________ l.kehoe@ieee.org Southeast (recruitment) Thomas M. Flynn Phone: +1 770 645 2944 Fax: +1 770 993 4423 Email: ______________ flynntom@mindspring.com Midwest (product) Dave Jones Phone: +1 708 442 5633 Fax: +1 708 442 7620 Email: ____________ dj.ieeemedia@ieee.org Will Hamilton Phone: +1 269 381 2156 Fax: +1 269 381 2556 Email: wh.ieeemedia@ieee.org ____________ Joe DiNardo Phone: +1 440 248 2456 Fax: +1 440 248 2594 Email: ____________ jd.ieeemedia@ieee.org Connecticut (product) Stan Greenfield Phone: +1 203 938 2418 Fax: +1 203 938 3211 Email: greenco@optonline.net ____________ Southern CA (product) Marshall Rubin Phone: +1 818 888 2407 Fax: +1 818 888 4907 Email: ____________ mr.ieeemedia@ieee.org Northwest/Southern CA (recruitment) Tim Matteson Phone: +1 310 836 4064 Fax: +1 310 836 4067 Email: ____________ tm.ieeemedia@ieee.org Southeast (product) Bill Holland Phone: +1 770 435 6549 Fax: +1 770 435 0243 Email: ____________ hollandwfh@yahoo.com Japan Tim Matteson Phone: +1 310 836 4064 Fax: +1 310 836 4067 Email: ____________ tm.ieeemedia@ieee.org Europe (product/recruitment) Hillary Turnbull Phone: +44 (0) 1875 825700 Fax: +44 (0) 1875 825701 Email: _______________ impress@impressmedia.com Boldface denotes advertisements in this issue. Advertising Personnel Computer IEEE Computer Society 10662 Los Vaqueros Circle Los Alamitos, California 90720-1314 USA Phone: +1 714 821 8380 Fax: +1 714 821 4010 http://www.computer.org advertising@computer.org __________________ Computer Marion Delaney IEEE Media, Advertising Director Phone: +1 415 863 4717 Email: _____________ md.ieeemedia@ieee.org Marian Anderson Advertising Coordinator Phone: +1 714 821 8380 Fax: +1 714 821 4010 Email: ______________ manderson@computer.org Sandy Brown IEEE Computer Society, Business Development Manager Phone: +1 714 821 8380 Fax: +1 714 821 4010 Email: sb.ieeemedia@ieee.org ____________ Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F CAREER OPPORTUNITIES CAREER OPPORTUNITIES MICHIGAN TECH, Computer Engineering – Senior Faculty Position. The Department of Electrical and Computer Engineering at Michigan Technological University invites applications for a tenured faculty position in computer engineering. The department is seeking an established researcher in real-time computing. Areas of particular interest include real-time hardware, RTOS, and the design and implementation of embedded and/or distributed real-time systems. We are looking for a person whose central focus in real-time systems can provide technical leadership and help integrate several existing research projects involving peripheral aspects of realtime computing. A successful senior candidate will have a demonstrated track record of establishing and conducting a high quality research program, sufficient to qualify for the position of Associate or Full Professor with tenure. A candidate with a demonstrated potential for establishing a high quality research program will be considered for a position of Assistant Professor. The Department of Electrical and Computer Engineering has one of the leading undergraduate programs in the US and is aggressively growing its graduate and research programs. Michigan Tech is located in the beautiful Upper Peninsula of Michigan, offering extensive outdoor recreation. Michigan Tech is an Philips has the following job opportunities available (various levels/types): PHILIPS ELECTRONICS NORTH AMERICA ANDOVER, MA • Software Engineer (SWE-PENAC-MA) • Project Manager (PM-PENAC-MA) – Manage and coordinate design, development and implementation activities for the various technology projects FRAMINGHAM, MA • Software Engineer (SWE-PENAC-F-MA) FOSTER CITY, CA • Software Engineer (SWE-PENAC-CA) • Test Engineer (TE-PENAC-CA) EL PASO,TX • Engineering Manager (EM-PENAC-TX) PHILIPS MEDICAL SYSTEMS MR LATHAN, NY • Mechanical Engineer (MEN-PMR-NY) • Manufacturing Engineer (ME-PMR-NY) PHILIPS NUCLEAR MEDICINE FITCHBURG, WI • Software Engineer (SWE-PNM-WI) MILPITAS, CA • Software Engineer (SWE-PNM-CA) PHILIPS ULTRASOUND BOTHELL, WA • Software Engineer (SWE-PU-WA) Some positions may require travel. Submit resume by mail to PO box 4104, Santa Clara, CA 950564104, ATTN: HR Coordinator. Must reference job title and job code (i.e. SWE-PENAC-CA) in order to be considered. EOE. 86 Computer equal opportunity employer. Send resume, statements of teaching and research interests, and contact data for three references to ____________ cpesearch@mtu.edu. IOWA STATE UNIVERSITY, Electrical and Computer Engineering Department. The Electrical and Computer Engineering Department at Iowa State University has immediate openings for faculty positions at all levels. Applications will be accepted from highly qualified individuals for regular faculty positions in the department in all core areas of expertise in Electrical or Computer Engineering, especially in •Computer engineering with emphasis on embedded systems; •VLSI with emphasis on analog/mixedsignal/RF IC design and bio applications; •Software engineering; •Information assurance and security; and •Distributed decisions sciences, controls, and applications. Faculty positions are also available in Interdisciplinary research areas as part the Iowa State University College of Engineering’s aggressive mission to fill 50 new college-wide positions with faculty who possess the talent to address the challenges that define worldwide quality of life and have global impact. The positions are targeted in the following interdisciplinary research and education cluster areas: •Biosciences and Engineering, •Energy Sciences and Technology, •Engineering for Extreme Events, •Information and Decision Sciences, •Engineering for Sustainability. Duties for all positions will include undergraduate and graduate education, developing and sustaining externally-funded research, graduate student supervision and mentoring, and professional/institutional service. All candidates must have an earned Ph.D. degree in Electrical Engineering, Computer Engineering, Computer Science or related field; and the potential to excel in the classroom and to establish and maintain a productive externally funded research program. Associate and Full Professor candidates must, in addition, have an excellent record of externally funded research and internationally recognized scholarship. Exceptional senior candidates may be considered for endowed research chair/professorship positions. Rank and salary are commensurate with qualifications. Screening will begin on November 1, 2007, and will continue until positions are filled. To guarantee consideration, complete applications must be received by 1/18/2008. For regular faculty positions, apply online at http://iastate.jobs.com. Vacancy #070478. For information on positions in the cluster areas and application process, visit http://www.engineering.iastate. edu/clusters. ISU is an EO/AA employer. _______ UNIVERSITY OF CALIFORNIA, SANTA CRUZ, Computer Engineering, Associate/Full Professor, Autonomous Systems. The Computer Engineering Department invites applications for a tenured (Associate or Full Professor) position in Autonomous Systems. Potential areas of specialization include robotics, control, mechatronics, and assistive technology. The department is launching an initiative in autonomous systems and mechatronic engineering, and seeks an individual to join our core faculty in this area and lead the development of new research and degree programs. Apply: http://www.soe.ucsc.edu/jobs, Position #808. To ensure full consideration, applications must arrive by Jan. 1, 2008. EEO/ AA/IRCA Employer. UNIVERSITY OF NORTH CAROLINA WILMINGTON, Computer Science (Assistant Professor, Tenure-track). August 2008. Ph.D. (or ABD) in Computer Science or closely related area required. Emphasis in computer and network security or closely related area. SUBMISSION DETAILS: Rates are $299.00 per column inch ($320 minimum). Eight lines per column inch and average five typeset words per line. Send copy at least one month prior to publication date to: Marian Anderson, Classified Advertising, Computer Magazine, 10662 Los Vaqueros Circle, PO Box 3014, Los Alamitos, CA 90720-1314; (714) 821-8380; fax (714) 821-4010. Email: _________________ manderson@computer.org. In order to conform to the Age Discrimination in Employment Act and to discourage age discrimination, Computer may reject any advertisement containing any of these phrases or similar ones: “…recent college grads…,” “…14 years maximum experience…,” “…up to 5 years experience,” or “…10 years maximum experience.” Computer reserves the right to append to any advertisement without specific notice to the advertiser. Experience ranges are suggested minimum requirements, not maximums. Computer assumes that since advertisers have been notified of this policy in advance, they agree that any experience requirements, whether stated as ranges or otherwise, will be construed by the reader as minimum requirements only. Computer encourages employers to offer salaries that are competitive, but occasionally a salary may be offered that is significantly below currently acceptable levels. In such cases the reader may wish to inquire of the employer whether extenuating circumstances apply. Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F COLLEGE OF ENGINEERING ALFAISAL UNIVERSITY Riyadh, Saudi Arabia Alfaisal University is a private, not-for-profit, research university, comprising the Colleges of Engineering, Science and General Studies, Medicine, and Business, and will commence its programs in Fall 2008. The language of instruction is English and modern learning outcomes, paradigms and technologies are used. The university was founded by King Faisal Foundation along with organizations such as Boeing, British Aerospace, Thales, and King Faisal Specialist Hospital & Research Center, who serve on its Board of Trustees. The College of Engineering will offer undergraduate and graduate programs in the following disciplines and subdisciplines: ELECTRICAL (power, communications, signal processing, electronics, photonics), COMPUTER (intelligent systems, language and speech, computer systems, computation), MECHANICAL (applied mechanics, product creation), AEROSPACE (thermo/fluid systems, aerospace systems, transportation, system dynamics and control), MATERIALS (materials processing, materials properties and performance, polymers, nanoscience and technology), CHEMICAL (catalysis, reactor design, design-systems, polymers). All programs have been developed by renowned scholars from leading universities in the US and the UK, and are designed to be qualified for accreditation according to US and UK standards and requirements. Alfaisal Engineering seeks candidates for the following positions, commencing in August 2008: FOUNDING SENIOR FACULTY (with instructional, research, and administrative responsibilities), RESEARCH SCIENTISTS (academics with research focus), LECTURERS (academics with instructional focus), POST-DOCS (Doctorate degree holders with research focus), INSTRUCTORS (Masters degree holders with instructional focus), and ENGINEERS (Bachelors degree holders). Attractive salary and start-up support is provided. Queries and Applications should be sent to engnr_recruiting@alfaisal.edu. ______________________ The subject line should specify the discipline, subdiscipline, position, and the advertisement reference. The deadline for applications is 31 December 2007. Interviews for leading positions will be conducted in January and February 2008 in Cambridge, MA, USA, and Cambridge, England, UK. ______________ __________ ________________ ________________ 87 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Details at __________________ http://www.uncw.edu/csc/posi_____ Screening begins January 9, tion.htm. 2008. EEO/AA Employer. Women and minorities encouraged to apply. ______________ _________________ _____________ ARIZONA STATE UNIVERSITY School of Computing and Informatics Department of Computer Science and Engineering The Department of Computer Science and Engineering has a Lecturer position open. Applicants are required to have their MS in computer science, computer engineering, or a closely related field. Applicants who have completed two years postmasters or are currently PhD candidates in their programs will also be considered. In addition they must show evidence of effective written, communication and management skills and a willingness to learn and manage course servers. A Ph.D. in computer science or a related field and teaching experience at a university, college, or junior college level is desired. The successful candidates will be expected to teach and co-ordinate several on-line sections of a computer literacy course (CSE 180) and teach basic programming courses in C# (CSE 182) and other languages. It is expected that the teaching will be through various delivery methods including on-line and other distance learning formats. Depending on the course load, the candidate is expected to provide service on department committees and develop new curriculum for both existing degree programs as well as for new programs. Applications will be reviewed as received until the search is closed. Early applications are strongly encouraged. Application packages must include a cover letter, detailed curriculum vitae, teaching statement, and the names, addresses, and phone numbers of three professional references. Application packages must be uploaded via http://sci.asu.edu/hiring. Please direct any questions to the Chairperson of the recruiting committee at ______________ cse.recruiting@asu.edu. The closing date for receipt of applications is January 15, 2008. If not filled, applications will be reviewed on the 15th and 30th of the month thereafter until the search is closed. Anticipated start date is August 16, 2008. Arizona State University is an Equal Opportunity/Affirmative Action Employer. A Background check is required for employment. 88 Computer UNIVERSITY OF CALIFORNIA, LOS ANGELES, Department of Computer Science. The Department of Computer Science in the Henry Samueli School of Engineering and Applied Science at the University of California, Los Angeles, invites applications for tenuretrack positions in all areas of Computer Science and Computer Engineering. Applications are also encouraged from distinguished candidates at senior levels. Quality is our key criterion for applicant selection. Applicants should have a strong commitment to both research and teaching and an outstanding record of research for their level. We seek applicants in any mainstream area of Computer Science and Computer Engineering such as software systems, embedded systems and machine learning as well as those with a strength in emerging technologies related to computer science such as biocomputing, nano architectures, and nanosystems. To apply, please visit http://www.cs.ucla.edu/recruit. Faculty applications received by January 31 will be given full consideration. The University of California is an Equal Opportunity/Affirmative Action Employer. WASHINGTON UNIVERSITY IN SAINT LOUIS, Department of Computer Science and Engineering, Faculty Positions. The School of Engineering and Applied Science at Washington University has embarked on a major initiative to expand its programs and facilities. As part of this initiative, the Department of Computer Science and Engineering is seeking outstanding faculty in the broad area of digital systems and architecture, including embedded computing, advanced multi-core architectures and hybrid computing systems. We have a special interest in candidates seeking to develop multi-disciplinary collaborations with colleagues in related disciplines. On the applications side, this may include collaborations in systems biology, neural engineering and genetics. On the technology and basic science side, it may include collaborations in electrical engineering, materials and physics. Successful candidates must show exceptional promise for research leadership and have a strong commitment to high quality teaching at all levels. Our faculty is engaged in a broad range of research activities including hybrid computing architectures, networking, computational biology, robotics, graphics, computer vision, and advanced combinatorial optimization. The department provides a supportive environment for research and the preparation of doctoral students for careers in research. Our doctoral gradu- Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page ates go on to positions of leadership in both academia and industry. The department values both fundamental research and systems research with the potential for high impact, and has a strong tradition of successful technology transfer. Limits on undergraduate enrollments and the university’s growing popularity allow us to offer small classes and close personal attention to a diverse student body of exceptional quality. A faculty known for its collegiality provides a supportive environment for new arrivals. A progressive administration reaches out to academic couples seeking to co-locate, and promotes policies that reward research, teaching, and innovative new initiatives. Washington University is one of the nation’s leading research universities and attracts top-ranked students from across the country and around the world. It is a medium-sized institution, with roughly 6,000 full-time undergraduates and 6,000 graduate students, allowing it to provide both a strong sense of community and a broad range of academic opportunities. Its six professional schools provide advanced education in engineering, medicine, social work, business, law, architecture and art. It has exceptional research strengths in the life sciences and medicine, creating unmatched opportunities for interdisciplinary collaborations for faculty in computer science and engineering. It has one of the most attractive university campuses anywhere, and is located in a lovely residential neighborhood, adjacent to one of the nation’s largest urban parks, in the heart of a vibrant metropolitan area. St. Louis is a wonderful place to live, providing access to a wealth of cultural and entertainment opportunities, while being relatively free of the everyday hassles and pressures of larger cities. Applicants should hold a doctorate in Computer Engineering, Computer Science or Electrical Engineering. Qualified applicants should submit a complete application (cover letter, curriculum vita, research statement, teaching statement, and names of at least three references) electronically to ____ recruiting@cse.wustl.edu. ___________ Other communications may be directed to Dr. Jonathan Turner, _____________ jon.turner@wustl.edu. Applications will be considered as they are received. Applications received after January 15, 2008 will receive limited consideration. Washington University is an equal opportunity/affirmative action employer. UNIVERSITY OF PORTLAND, Faculty Position in Computer Science. The Department of Electrical Engineering and Computer Science at the University of A BEMaGS F Portland seeks a computer science faculty member for a tenure-track position at the assistant or associate professor level, to begin August 2008. Ph.D. is required; dedication to excellence in teaching is essential. Must be a U.S. citizen or permanent resident. Areas of interest include computer security, computer graphics, computer gaming, robotics, and computer systems. Other areas of expertise will be considered. Duties include undergraduate teaching and advising, laboratory development, engagement in scholarly activity, and service to the University. A typical teaching responsibility is three courses per semester. The University of Portland is an independently governed Catholic university that welcomes people of diverse backgrounds. The University serves approximately 3500 students. The CS and EE programs are both ABETaccredited. Send hard copy of application materials (letter of interest, vita, teaching statement, research statement, references) to: CS Search Committee, School of Engineering, University of Portland, 5000 N. Willamette Blvd., Portland, OR 97203 or apply via email at cssearch@up. _______ edu. Applications will be processed as __ they arrive. For full consideration, please apply by December 1. The University of Portland is an affirmative action/equal opportunity employer. For information UNIVERSITY OF TORONTO The Edward S. Rogers Sr. Department of Electrical & Computer Engineering 10 King’s College Road Toronto, Ontario, Canada M5S 3G4 The Edward S. Rogers Sr. Department of Electrical and Computer Engineering at the University of Toronto invites applications for tenure-stream Assistant or Associate Professor positions, beginning July 1, 2008, in two areas: 1. Computer or FPGA Architecture. Research areas of interest include, but are not limited to: multi- or single- core processor architecture, FPGA architecture and CAD, embedded processor design, programming and compiler support for multi-core and novel processors, memory systems, programmable architectures for integrated digital circuits, and power-aware and power efficient architectures. Applications and references for this position should be addressed to Professor Tarek Abdelrahman, Chair of Search Committee and sent to: CompFPGASearch@ece.utoronto.ca. ______________________ 2. Information Security. Research areas of interest include identity, privacy and security information technologies for: computer networks, distributed systems, sensor and networked systems, embedded systems, computer architecture and system survivability. Applications and references for this position should be addressed to Professor Dimtrios Hatzinakos, Chair of Search Committee and sent to: InfoSecSearch@ece.utoronto.ca. ____________________ Candidates must have (or are about to receive) a Ph.D. in the relevant area. The department ranks among the top 10 ECE departments in North America. It attracts outstanding students, has excellent facilities, and is ideally located in the middle of a vibrant, artistic, and diverse cosmopolitan city. The department offers highly competitive salaries and start-up funding, and faculty have access to significant Canadian research operational and infrastructure grants. Additional information can be found at: www.ece.utoronto.ca. The successful candidates are expected to pursue excellence in research and teaching at both the graduate and undergraduate levels. Applicants must submit their application by electronic email to one of the two email addresses given above. Please submit only Adobe Acrobat PDF documents. Applicants will receive an email acknowledgement. All applications should include: a curriculum vitae, a summary of previous research and proposed new directions, and a statement of teaching philosophy and interests. In addition, applicants must arrange to have three confidential letters of recommendation sent directly (by the referee) by email to the correct address given above. Applications and referee-sent references should be received by January 15, 2008. The University of Toronto is strongly committed to diversity within its community and especially welcomes applications from visible minority group members, women, Aboriginal persons, persons with disabilities, members of sexual minority groups, and others who may contribute to the further diversification of ideas. All qualified candidates are encouraged to apply; however, priority will be given to Canadian Citizens and Permanent Residents. 89 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page about the University and program refer to http://orgs.up.edu/cssearch. Direct additional inquiries about the position to cssearch@up.edu or 503-943-7314. __________ FLORIDA INTERNATIONAL UNIVERSITY, School of Computing and Information Sciences. Applications are invited for multiple tenure-track or tenured faculty positions at the levels of Assistant, Associate, or Full Professor. A Ph.D. in Computer Science or related areas is required. Outstanding candi- dates are sought in areas of (1) Software and Computer Security; (2) Software and Computer Systems; (3) Bio/Medical/ Health Informatics; (4) Data Mining; and (5) Human-Computer Interface (HCI). Exceptional candidates in other areas will be considered as well. Candidates with the ability to forge interdisciplinary collaborations will be favored. Candidates for senior positions must have a proven record of excellence in research funding, publications, teaching, and professional service, as well as demonstrated ability for developing and leading collaborative CHAIR Computer Science and Engineering University of Michigan, Ann Arbor The newly restructured Division of Computer Science and Engineering (CSE) in the College of Engineering at the University of Michigan is seeking an inaugural Chair. In recognition of its increasing prominence in the College’s broader education and research missions, the Division’s administrative autonomy and profile have recently been raised to match other College departments, providing the new Chair with an unprecedented opportunity to lead a strong CSE program into a new era. Computer Science at Michigan will soon celebrate its 50th anniversary, and the Computer Engineering program is among the oldest in the nation. CSE is currently the third largest academic unit in the College, with over 40 faculty and about 200 graduate and 325 undergraduate students. The state-of the-art 104,000 square foot CSE Building opened in 2006, housing CSE faculty and projects that had previously been distributed across several buildings under one roof. For more CSE details, see www.cse.umich.edu. CSE has a strong commitment to excellence in computing education, including interdisciplinary curricula with other units in the University. With external funding of approximately $12 million annually, CSE is a world-class leader in core research areas such as computer architecture, software systems and artificial intelligence. CSE has close and longstanding research and educational collaborations with the Division of Electrical and Computer Engineering (ECE), and these are preserved in the restructured organization by having the Chairs of CSE and ECE serve as co-Chairs of the Electrical Engineering and Computer Science (EECS) Department. More broadly, CSE teams with colleagues from across the University to spearhead interdisciplinary investigations in areas like bioinformatics, computational economics, robotics and educational software. CSE is poised to use its new autonomy to shape research and educational innovations and collaborations that respond to and anticipate the emerging roles of computation in our society. In this time of transition and opportunity for CSE, a successful Chair candidate must have outstanding leadership, collaborative and administrative abilities, and be an outstanding scholar. The candidate should possess a compelling vision for the future of CSE research and education, should be able to work with a diverse group of faculty, staff, students and administrators to articulate and achieve common goals in pursuit of this vision and should be able to marshal support from alumni and industry in these endeavors. He or she should have an exemplary record of achievement in research, teaching and service commensurate with appointment as a tenured full professor. Inquiries, nominations and applications (with CV) should be sent to ____________ or to Prof. Edmund Durfee, CSE Chair Search CSEChair@umich.edu, Advisory Committee, 3633 CSE Building, University of Michigan, Ann Arbor, MI, 48109-2121. The University of Michigan is an Equal Opportunity Affirmative Action Employer. Individuals from under-represented groups are encouraged to apply. 90 Computer A BEMaGS F research projects. Outstanding candidates for the senior positions will be considered for the endowed Ryder Professorship position. Successful candidates are expected to develop a high-quality funded research program and must be committed to excellence in teaching at both graduate and undergraduate levels. Florida International University (FIU), the state university of Florida in Miami, is ranked by the Carnegie Foundation as a comprehensive doctoral research university with high research activity. FIU offers over 200 baccalaureate, masters and doctoral degree programs in 21 colleges and schools. With over 38,000 students, it is one of the 25 largest universities in the United States, and boasts a new and accredited Law School and the newly created College of Medicine. US News & World Report has ranked FIU among the top 100 public universities, and Kiplinger’s Personal Finance magazine ranked FIU among the best values in public higher education in the country in their 2006 survey. The School of Computing and Information Sciences (SCIS) is a rapidly growing program of excellence at the University. The School has 31 faculty members (including seven new faculty members hired in the last three years), 1,200 students, and offers B.S., M.S., and Ph.D. degrees in Computer Science and B.S. and B.A. degrees in Information Technology. Its undergraduate program is the largest among the ten state universities in Florida and SCIS is the largest producer of Hispanic CS and IT graduates in the US. The Ph.D. enrollment in the School has doubled in the last four years with around 80 enrolled Ph.D. students. In 2006-07, the School received $2.7M in competitive research grants and leads the similar programs in the State of Florida in terms of per faculty annual research funding. In addition, the school receives an annual average of $2.2M of in-kind grants and donations from industry. Its research has been sponsored by NSF, NASA, NIH, ARO, ONR, NOAA, and other federal agencies. Several new faculty members have received the prestigious NSF CAREER AWARD, DoE CAREER AWARD, and IBM Faculty Research Awards. SCIS has broad and dynamic partnerships with industry. Its research groups include the NSF CREST Center for Emerging Technologies for Advanced Information Processing and High-Confidence Systems, the High Performance Database Research Center, the Center for Advanced Distributed Systems Engineering, the IBM Center for Autonomic and Grid Computing, and other research laboratories. The SCIS has excellent computing infrastructure and technology support. In addition, the SCIS faculty and students have access to the grid computing infrastructure with 1000 nodes under the Latin American Grid (LA Grid) Consortium (http://lagrid.fiu.edu), a first ever comprehensive international part- Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page nership, co-founded by IBM and FIU, linking institutions in the US, Latin America, and Spain for collaborative research, innovation and workforce development. Applications, including a letter of interest, contact information, curriculum vitae, and the names of at least three references, should be sent to Chair of Recruitment Committee, School of Computing and Information Sciences, Florida International University, University Park, Miami, FL 33199. E-mail submission to recruit@cis.fiu.edu ___________ is preferred. The application review process will begin on January 15, 2008, and will continue until the positions are filled. Further information can be obtained from the School website http://www.cis.fiu.edu, or by email to ___________ recruit@cis.fiu.edu. Women and members of underrepresented groups are strongly encouraged to apply. Florida International University is a member of the State University System of Florida and is an equal opportunity/affirmative action/equal access employer. PURDUE UNIVERSITY. The Department of Computer Science at Purdue University invites applications for tenure-track positions beginning August 2008. While outstanding candidates in all areas of Computer Science will be considered, preference will be given to applicants with a demonstrable research record in operating systems, software engineering, and theory. Candidates with a multi-disciplinary focus are also encouraged to apply. Of special interest are applicants with research focus in computational science and engineering, bioinformatics, and health-care engineering. The level of the positions will depend on the candidate’s experience. The Department of Computer Science offers a stimulating and nurturing academic environment. Forty-four faculty members direct research programs in analysis of algorithms, bioinformatics, databases, distributed and parallel computing, graphics and visualization, information security, machine learning, networking, programming languages and compilers, scientific computing, and software engineering. The department has implemented a strategic plan for future growth supported by the higher administration and recently moved into a new building. Further information about the department and more detailed descriptions of the open positions are available at ____ http:// www.cs.purdue.edu. Information about the multi-disciplinary hiring effort can be found at http://www.science.purdue. __________ All applicants should edu/COALESCE/. hold a PhD in Computer Science, or a closely related discipline, be committed to excellence in teaching, and have demonstrated potential for excellence in research. Salary and benefits are highly competitive. Applicants are strongly encouraged to apply electronically by sending their curriculum vitae, research A BEMaGS F College of Engineering and Mineral Resources LANE DEPARTMENT OF COMPUTER SCIENCE AND ELECTRICAL ENGINEERING Tenure Track Faculty Position in Computer Networks/Security The Lane Department of Computer Science and Electrical Engineering at West Virginia University (WVU) anticipates filling one or more tenure-track faculty positions at the Assistant Professor level, subject to availability of funds. The successful applicant will be expected to teach graduate and undergraduate coursework and develop an externally sponsored research program in the areas of computer systems, networks and computer security. Applicants must have an earned Ph.D. in Computer Science, Computer Engineering, or a closely related discipline at the time of appointment, and must have demonstrated ability or potential to contribute to the research and teaching missions of the department, including the ability to teach a broad range of undergraduate topics in Computer Science. WVU is a comprehensive land grant institution of over 27,000 students with medical, law, and business schools. The department has 31 tenure-track faculty members, 400 undergraduates, and 280 graduate students. It offers BS degrees in Computer Science, Computer Engineering, Electrical Engineering and Biometric Systems; MS degrees in Computer Science, Software Engineering, and Electrical Engineering; and Ph.D. degrees in Computer Science, Computer Engineering and Electrical Engineering. The department conducts approximately $5 million annually in externally sponsored research, with major research activities in the areas of biometric systems, bioengineering, information assurance, nanotechnology, power systems, software engineering, virtual environments, and wireless networks. Strong opportunities exist for building collaborative partnerships with nearby federal research facilities, including the Department of Defense, Department of Energy, NASA, and the FBI. Review of applications will begin January 1, 2008 and continue until positions are filled. Women and minorities are strongly encouraged to apply. Applicants should send a letter describing their qualifications, a curriculum vita, statements of teaching philosophy and research objectives, and names and e-mail addresses of at least 3 references to: Dr. Bojan Cukic, Faculty Search Committee Chair Lane Department of Computer Science and Electrical Engineering P.O. Box 6109 West Virginia University Morgantown, WV 26506-6109 Telephone: 304-293-LANE (x 2526) Email: _____________ cs-search@mail.wvu.edu web: http://www.csee.wvu.edu West Virginia University is an equal opportunity/affirmative action employer. Careers with Mass Appeal Assistant Professor Department of Computer Science www.cs.umb.edu The Computer Science Department at the University of Massachusetts Boston invites applications for Fall 2008 for one faculty position at the Assistant Professor level. We offer a BS, an MS with an emphasis on software engineering, and a Ph.D. in computer science. We seek to strengthen our research program significantly. Current faculty interests include biodiversity informatics, computer and human vision, data mining, databases, networks, software engineering, system modeling, and theoretical computer science. Strong candidates will be considered from any area of Computer Science, but preference will be given to a candidate who does research in artificial intelligence, particularly evolutionary computing, knowledge representation, machine learning or neural networks. Evidence of significant research potential and a Ph.D. in computer science or a related area are required. We offer a competitive salary and a generous start-up package. Send cover letter, curriculum vitae, statements about research and teaching, and the names and email addresses of three references to Search 680E at search@cs.umb.edu. ___________ Our campus overlooks Boston harbor; our faculty and students enjoy professional life in a center of academia and the software industry. For more information, visit us at http://www.cs.umb.edu. Review of applications has begun and will continue until the position is filled. UMass Boston is an affirmative action, equal opportunity Title IX employer. 91 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page and teaching statements, and names and contact information of at least three references in PDF to ___________ fac-search@cs.purdue.edu. Hard copy applications can be _____ sent to: Faculty Search Chair, Department of Computer Science, 305 N. University Street, Purdue University, West Lafayette, IN 47907. Applicants matching one search may be considered in other relevant searches when appropriate. Review of applications will begin on October 1, 2007, and will continue until the positions are filled. Purdue University is an Equal Opportunity/Equal Access/Affirmative Action employer fully committed to achieving a diverse workforce. WAYNE STATE UNIVERSTY, Department of Computer Science, TenureTrack Faculty Position. The Department of Computer Science of Wayne State University invites applications for a tenure-track faculty position at the Assistant/Associate Professor level. Continuing our recent growth, we are seeking applicants in the areas of Software Engineering and Bioinformatics. Outstanding applicants in other areas will also be considered. Candidates should have a Ph.D. in computer science or a related area. The successful candidate will have a strong commitment to both research and teaching, a strong publication record and potential for obtaining external research funding. Senior applicants should have strong publication and funding records. Currently, the department has 19 faculty, 78 Ph.D. and 120 M.S. students. The Department’s total annual R&D expenditures average between $2-3 million in research areas, including bioinformatics, software engineering, systems, databases and image processing. Our junior faculty benefit from an extraordinarily supportive environment as demonstrated by their success in securing two recent NSF CAREER awards, as well as other very competitive research funding. Faculty actively collaborate with many other centers and departments, including the School of Medicine, which is the largest single-campus medical school in the country, and the Karmanos Cancer Institute, a nationally recognized comprehensive cancer center. More information about the department can be found at: http://www.cs.wayne.edu. Wayne State University is a premier institution of higher education offering more than 350 undergraduate and graduate academic programs to more than 33,000 students in 11 schools and colleges. Wayne State ranks in the top 50 nationally among public research universities. As Michigan’s only urban university, Wayne State fulfills a unique niche in providing access to a world-class education. The University A BEMaGS F offers excellent benefits and a competitive compensation package. Submit applications online at http://jobs.wayne. edu, __ refer to Posting # 034690 and include a letter of intent, a statement of research and teaching interests, a CV, and contact information for at least three references. All applications received prior to January 11, 2008 will receive full consideration. However, applications will be accepted until the position is filled. Wayne State University is an equal opportunity/affirmative action employer. THE HONG KONG POLYTECHNIC UNIVERSITY, Department of Computing.The Department invites applications for Assistant Professors in most areas of Computing, including but not limited to Software Engineering / Biometrics / Digital Entertainment / MIS and Pervasive Computing. Applicants should have a PhD degree in Computing or closely related fields, a strong commitment to excellence in teaching and research as well as a good research publication record. Initial appointment will be made on a fixed-term gratuity-bearing contract. Re-engagement thereafter is subject to mutual agreement. Remuneration package will be highly competitive. Applicants should state their current and expected salary in the application. Please submit UNIVERSITY OF BRIDGEPORT Engineering Faculty Positions Available Computer Science and Engineering DEPARTMENT OF ELECTRICAL & COMPUTER ENGINEERING Computer Engineering Faculty Position in High-Performance and Reconfigurable Computing The Department of Electrical and Computer Engineering at The George Washington University invites applications for tenuretrack, tenured and contractual non-tenure-accruing faculty positions at all ranks, in the area of Computer Engineering. Two positions will be for tenure-track/tenured faculty, and the third position will be a one-year renewable non-tenure-accruing contractual position at the Assistant/Associate Professor rank, and successful candidates may start as early as Spring 2008. Faculty with research in High-Performance Computing and Reconfigurable Computing are particularly encouraged to apply, however, all areas of Computer Engineering will be considered. Additional information and details on position qualifications and the application procedure are available on http://www.ece.gwu.edu. Review of applications will continue until the positions are filled. The George Washington University is an Equal Opportunity/Affirmative Action Employer 92 Computer The fast-growing department of Computer Science and Engineering at the University of Bridgeport invites applications for full time tenure-track positions at the Assistant/Associate Professor levels. Candidates for tenuretrack positions must have a Ph.D. in computer science, computer engineering, or a related field. A strong interest in teaching undergraduate and graduate courses and an excellent research record are required. The ability to teach lab-based courses is also required. Applicants are sought in the areas of Medical Electronics, Biomedical Engineering, Biometrics, Bio-Informatics, Wireless Design, Distributed Computing, Computer Architecture, Data Base Design, Algorithms, e-commerce, Data Mining and Artificial Intelligence. There are opportunities to participate in the external engineering programs, which include weekend and evening graduate and continuing education classes, on-site instruction in local industry and distance learning initiatives. Electrical Engineering The Department of Electrical Engineering at the University of Bridgeport is searching for two tenure track Assistant Professors to start January 7, 2008. Qualifications include an earned doctorate in Electrical Engineering or equivalent. All disciplines are encouraged to apply, but preference will be given (i) to those in the area of sensors: biological and chemical, including design and theoretical modeling of semiconductor sensors and thin films and (ii) to those involved with SOC (systems on a chip) design, digital/analog VLSI design and testing, and VLSI design automation. Other areas of engineering will be considered, including RFIC design, A/D wireless communications, Biomedical Engineering, and Medical Electronics. The successful candidate is expected to excel in teaching undergraduate and graduate students in the classroom and in the lab. A strong publication/ research record is required as well. Applicants should send a cover letter, resume and address and e-mail address of four references via e-mail to: University of Bridgeport, School of Engineering Faculty Search Committee: _____________ enfacrec@bridgeport.edu The University of Bridgeport is an Affirmative Action/Equal Opportunity Employer, we encourage women and minorities to apply Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page _____ your application via email to hrstaff@ polyu.edu.hk. ________ Application forms can be downloaded from http://www.polyu. edu.hk/hro/job.htm. ____________ Recruitment will continue until the positions are filled. Details of the University’s Personal Information Collection Statement for recruitment can be found at http://www.polyu. edu.hk/hro/jobpics.htm. ______________ McMURRY UNIVERSITY. Assistant Professor of Computer Science to begin August 2008. Candidates must have Computer Science PhD. Position requires teaching 12 hours per semester, advising CS majors, and system admin of CS server. McMurry University is a private liberal arts institution affiliated with the United Methodist Church. Candidates must be willing to support the university mission and values. Review of applications will begin immediately and continue until position is filled. Complete details available at www.mcm.edu/employment. UNIVERSITY OF LOUISIANA AT LAFAYETTE, The Center for Advanced Computer Studies, Faculty Positions, Graduate Fellowships. Candidates with a strong research record and an earned doctorate in computer science or computer engineering are invited to apply for multiple tenure-track assistant/associate professor faculty positions starting fall of 2008. Target areas include Grid Computing, Large Scale Data & Knowledge Engineering, Distributed Software Systems, and Entertainment Computing. Consideration will also be given to outstanding candidates in other areas. Candidates must have demonstrated potential to achieve national visibility through accomplishments in research contract and graduate students. Faculty teach mostly at the graduate and senior undergraduate levels and offer a continuing research seminar. State and university funds are available to support research initiation efforts. Salaries are competitive along with excellent support directed towards the attainment of our faculty's professional goals. The Center's colloquium series brings many world known professionals to our campus each year. The Center is primarily a graduate research unit of 17 tenure-track and 7 research faculty, with programs leading to MS/PhD degrees in computer science and computer engineering. More than 200 graduate students are enrolled in these programs. The Center has been ranked 57th in a recent NSF survey based on research and development expenditures. The Center has state-of-the-art research and instructional computing facilities, consisting of several networks of SUN workstations and other high performance computing platforms. In addition, the Center has dedicated research laboratories in Intelligent Systems, Computer Architecture and Networking, Cryptog- raphy, FPGA and Reconfigurable Computing, Internet Computing, Virtual Reality, Entertainment Computing, Software Research, VLSI and SoC, Wireless Technologies, and Distributed Embedded Computing Systems. Related university programs include the CSAB (ABET) accredited undergraduate program in Computer Science, and the ABET accredited undergraduate program in Electrical and Computer Engineering. Additional information about the Center may be obtained at http://www.cacs.louisiana. edu. __ A number of PhD fellowships, val- A BEMaGS F ued at up to $24,000 per year including tuition and most fees, are available. They provide support for up to four years of study towards the PhD in computer science or computer engineering. Eligible candidates must be U.S. citizens or must have earned an MS degree from a U.S. university. Recipients also receive preference for low-cost campus housing. Applications may be obtained and submitted at http://gradschool.louisiana.edu. The University of Louisiana at Lafayette is a Carnegie Research University with high research activity, with an enrollment of THE INSTITUTE OF TECHNOLOGY AT THE UNIVERSITY OF WASHINGTON TACOMA Tenure-Track Faculty Positions The Institute of Technology at the University of Washington Tacoma is accepting applications for full-time, tenure-track faculty positions in the areas of Computing and Software Systems (CSS), Computer Engineering and Systems (CES), and Information Technology and Systems (ITS) beginning September 16, 2008. Commitment to high-quality teaching and to an externally funded research program are essential and excellent communication skills are required. Please visit our website for full descriptions of each position: http://www.tacoma.washing_________________ ton.edu/hr/jobs/ _________ COMPUTING AND SOFTWARE SYSTEMS CSS candidates with strengths in hardware-oriented CS (architecture and OS) or theoretical foundations will be given priority consideration, but we will consider all areas of expertise, with appropriate evidence of ability to teach core CS topics. The search is focused at the Assistant level but advanced candidates may be considered at the Associate and Full Professor levels. A Ph.D. in Computer Science or a closely related discipline, and an appropriately strong record of published research in computing is required. For additional information, please contact Dr. George Mobus at gmobus@u.washington.edu. ________________ COMPUTER ENGINEERING AND SYSTEMS CES candidates should have a strong background in a focused area such as: embedded and real-time systems, digital system design, or network design and security. The search is focused at the Assistant or Associate Professor levels but advanced candidates may be considered at the Professor level. A Ph.D. in Computing Engineering and Systems or a closely related field is required. For additional information, contact Dr. Larry Wear at lwear@u.washington.edu ______________ or by phone at (253) 692-4538. INFORMATION TECHNOLOGY AND SYSTEMS ITS candidates with applied research interests in asynchronous architectures/grid computing, graphics and game development, Web services and distributed computing, social networks and computing, collaborative learning technologies, databases, interaction and design, or information assurance & computer security are particularly encouraged to apply. Candidates will be considered at the ranks of Assistant or Associate Professor. A Ph.D. in Computer/Information Science, Information Technology, Information Systems or a closely related field is required by the time of appointment. Priority will be given to candidates with a background in industry or experience conducting/managing research programs outside academia. For additional information, please contact Dr. Orlando Baiocchi at baiocchi@u.washington.edu ________________ or by phone at (253) 692-4727. The University of Washington Tacoma, one of three UW campuses, is an urban undergraduate and master’s level campus that is changing the face of its region economically, culturally, and architecturally. The Institute of Technology was created in 2001 by a public/private partnership to address the critical industry demand for baccalaureate and masters level computing and engineering professionals. Please see http://www.insttech.washington.edu/ for additional information. Applications should include (1) a letter describing academic qualifications, (2) a statement of research interests, (3) a description of teaching philosophy, and (4) curriculum vitae. Applications must include contact information for at least three (3) references. Applications should be submitted electronically to: ________________ Ifaculty@u.washington.edu. Screening of credentials will begin November 19th. All positions will remain open until filled. Salary is competitive and will be commensurate with experience and qualifications. The University of Washington Tacoma is an affirmative action, equal opportunity employer. The University is committed to building a culturally diverse faculty and staff, and strongly encourages applications from women, minorities, individuals with disabilities and covered veterans. University of Washington Tacoma faculty engage in teaching, research and service and are expected to participate in lower division teaching. 93 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page over 16,000 students. Additional information may be obtained at http://www. louisiana.edu/. ________ The University is located in Lafayette, the hub of Acadiana, which is characterized by its Cajun music and food and joie de vivre atmosphere. The city, with its population of over 120,000, provides many recreational and cultural opportunities. Lafayette is located approximately 120 miles west of New Orleans. The search committee will review applications and continue until the positions are filled. Candidates should send a letter of intent, curriculum vitae, statement of research and teaching interests, and names, addresses and telephone numbers of at least four references. Additional materials, of the candidate's choice, may also be sent to: Dr. Magdy A. Bayoumi, Director, The Center for Advanced Computer Studies, University of Louisiana at Lafayette, Lafayette, LA 70504-4330. Tel: 337.482.6147. Fax: 337.482.5791. The University is an Affirmative Action/ Equal Opportunity Employer. SR. ORACLE DEVELOPER - EGB Systems , Stamford, CT - Develops DB applications & DB warehousing, DB modeling using TCA, OLAP, ORACLE & Oracle RDBMS. Bachelor in Science with 9 yrs exp in oracle, SQL and PL/SQL apply to careersusa@egbsystems.com. _________________ UNIVERSITY OF WASHINGTON, Computer Science & Engineering and Electrical Engineering, TenureTrack and Research Faculty. The University of Washington’s Department of Computer Science & Engineering and Department of Electrical Engineering have jointly formed a new UW Experimental Computer Engineering Lab (ExCEL). In support of this effort, the College of Engineering has committed to hiring several new faculty over the forth- The National University of Singapore (NUS) invites nominations and applications for the position of Head, Computer Science Department. NUS has about 23500 undergraduate and 9000 graduate students from 88 countries. In 2006, Newsweek's ranking of universities listed NUS as 31st globally and 3rd in Asia/Australasia. The CS Department has some 80 faculty members, many of whom regularly publish in prestigious conferences and journals, and serve on their program committees and editorial boards. We seek a new Head who can further raise the Department's achievements. The candidate should be an internationally-recognized researcher, with experience in technical leadership and team management. The salary and benefits are internationally competitive. For details, see: http://www.comp.nus.edu.sg/cs/________ cshodrec.html 94 Computer coming years. All positions will be dual appointments in both departments (with precise percentages as appropriate for the candidate). This year, we have two open positions, and encourage exceptional candidates in computer engineering, at tenure-track Assistant Professor, Associate Professor, or Professor, or Research Assistant Professor, Research Associate Professor, or Research Professor to apply. A moderate teaching and service load allows time for quality research and close involvement with students. The CSE and EE departments are co-located on campus, enabling cross department collaborations and initiatives. The Seattle area is particularly attractive given the presence of significant industrial research laboratories, a vibrant technology-driven entrepreneurial community, and spectacular natural beauty. Information about ExCEL can be found at <http://www.excel.wash______________ ington.edu>. We welcome applications in _______ all computer engineering areas including but not exclusively: atomics scale devices & nanotechnology, implantable and biologically-interfaced devices, synthetic molecular engineering, VLSI, embedded systems, sensor systems, parallel computing, network systems, and technology for the developing world. We expect candidates to have a strong commitment both to research and teaching. ExCEL is seeking individuals at all career levels, with appointments commensurate with the candidate’s qualifications and experience. Applicants for both tenure-track and research positions must have earned a PhD by the date of appointment. Please apply online at <http://www.excel.wash______________ ington.edu/jobs.html> with a letter of _____________ application, a complete curriculum vitae, statement of research and teaching interests, and the names of at least four references. Applications received by January 31st, 2008 will be given priority consideration. The University of Washington was awarded an Alfred P. Sloan Award for UNIVERSITY OF CALIFORNIA, RIVERSIDE The Department of Computer Science and Engineering invites applications for faculty positions at all levels and in all areas with special interest in interdisciplinary research involving Computer Graphics, Software Systems, High-Performance Computing and Computational Science and Engineering. Additional positions are available for candidates whose research interests intersect with the newly created Material Science and Engineering program (http://www.engr.ucr.edu/ mse/). __ Applicants must have a PhD in Computer Science or in a closely related field. Visit http://www.cs.ucr.edu for information and http://www.engr.ucr.edu/facultysearch/ to apply. Review of applications will begin on January 1, 2008, and will continue until the positions are filled. For inquiries contact __________ search@cs.ucr.edu. UC Riverside is an EqualOpportunity/Affirmative-Action Employer. A BEMaGS F Faculty Career Flexibility in 2006. In addition, the University of Washington is a recipient of a National Science Foundation ADVANCE Institutional Transformation Award to increase the participation of woman in academic science and engineering careers. We are building a culturally diverse faculty and encourage applications from women and minority candidates. STATE UNIVERSITY OF NEW YORK AT BINGHAMTON, Department of Computer Science, The Thomas J. Watson School of Engineering and Applied Science, http://www.cs. binghamton.edu. ____________ Applications are invited for a tenure-track position at the Assistant/Associate Professor level beginning in Fall 2008. Salary and startup packages are competitive. We are especially interested in candidates with specialization in (a) Embedded Systems and Compilers or (b) Ubiquitous Computing/Information Access or (c) Information Security or (d) Areas related to systems development. Applicants must have a Ph.D. in Computer Science or a closely related discipline by the time of appointment. Strong evidence of research capabilities and commitment to teaching are essential. We offer a significantly reduced teaching load for junior tenure track faculty for at least the first three years. Binghamton is one of the four Ph.D. granting University Centers within the SUNY system and is nationally recognized for its academic excellence. The Department has well- established Ph.D. and M.S. programs, an accredited B.S. program and is on a successful and aggressive recruitment plan. Local high-tech companies such as IBM, Lockheed-Martin, BAE and Universal Instruments provide opportunities for collaboration. Binghamton borders the scenic Finger Lakes region of New York. Submit a resume and the names of three references to the url address: __________________ http://binghamton.interviewexchange.com. _______ First consideration will be given to applications that are received by March 1, 2008. Applications will be considered until the positions are filled. Binghamton University is an equal opportunity/affirmative action employer. UNIVERSITY OF CHICAGO. The Department of Computer Science at the University of Chicago invites applications from exceptionally qualified candidates in all areas of Computer Science for faculty positions at the ranks of Professor, Associate Professor, Assistant Professor, and Instructor. The University of Chicago has the highest standards for scholarship and faculty quality, and encourages collaboration across disciplines. The Chicago metropolitan area provides a diverse and exciting environment. The local economy is vigorous, with international stature in banking, trade, commerce, manufactur- Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page ing, and transportation, while the cultural scene includes diverse cultures, vibrant theater, world-renowned symphony, opera, jazz, and blues. The University is located in Hyde Park, a pleasant Chicago neighborhood on the Lake Michigan shore. Please send applications or nominations to: Professor Stuart A. Kurtz, Chairman, Department of Computer Science, The University of Chicago, 1100 E. 58th Street, Ryerson Hall, Chicago, IL. 60637-1581 or to: apply-077714@mail____________ man.cs.uchicago.edu ____________ (attachments can be in pdf, postscript, or Word). Complete applications consist of (a) a curriculum vitae, including a list of publications, (b) forward-looking research and teaching statements. Complete applications for Assistant Professor and Instructor positions also require (c) three letters of recommendation, sent to ________ recommend077714@mailman.cs.uchicago.edu ____________________ or to the above postal address, including one that addresses teaching ability. Applicants must have completed, or will soon complete, a doctorate degree. We will begin screening applications on December 15, 2007. Screening will continue until all available positions are filled. The University of Chicago is an equal opportunity/affirmative action employer. DREXEL UNIVERSITY, Department of Computer Science, Faculty Positions. Drexel University's Department of Computer Science (www.cs.drexel.edu) invites applications for tenure-track faculty positions at all levels. The preferred interest is ARTIFICIAL INTELLIGENCE and MULTI-AGENT SYSTEMS, although exceptional applicants in other areas will be considered. The department has expanding graduate research and education programs in software engineering, graphics and vision, information assurance and security, human-computer interaction, high-performance computing and symbolic computation. We specialize in interdisciplinary and applied research and are supported by several major federal research grants from NSF, DARPA, ONR, DoD, DoE and NIST, as well as by private sources such as Intel, Nissan, NTT, and Lockheed Martin. The department offers BS, BA, MS, and Ph.D. degrees in computer science as well as BS and MS degrees in software engineering. Drexel is a designated National Security Agency (NSA) Center of Academic Excellence in Information Assurance Education. The department has over 425 undergraduate and over 100 graduate students, with annual research expenditures in excess of $4M. Several of the Computer Science faculty are recipients of NSF CAREER or Young Investigator Awards. Review of applications begins immediately. To assure consideration materials from applicants should be received by February 1, 2008. Successful applicants must demonstrate potential for research and teaching excellence in the environ- ment of a major research university. To be considered, please send an email to: cs__ search-08@cs.drexel.edu. ______________ Please include a cover letter, CV, brief statements describing your research program and teaching philosophy, and contact information for at least four references. Electronic submissions in PDF format are strongly preferred. THE UNIVERSITY OF TENNESSEE, KNOXVILLE. The Min Kao Department of Electrical Engineering and Computer Science (EECS) at The University of Tennessee, Knoxville is searching for candidates for tenuretrack faculty positions in all areas but preference will be given to applicants in computer engineering, including dependable and secure systems, wireless and sensor networks, embedded systems, and VLSI. Successful candidates must be committed to teaching at both undergraduate and graduate levels and have a strong commitment to research and a willingness to collaborate with other faculty in research. The Department currently enrolls approximately 400 undergraduate and about 250 graduate students. Faculty research expenditures currently average about $10 M per year. The department is starting a new growth phase thanks to gifts from alumnus Dr. Min Kao and other donors plus additional A BEMaGS F state funding totaling over $47.5 M for a new building and endowments for the department. The University of Tennessee and Battelle manage the nearby Oak Ridge National Laboratory, which provides further opportunities for research. Information about the EECS Department can be found at http://www.eecs. utk.edu/. Candidates should have an _____ earned Ph.D. in Electrical Engineering, Computer Engineering, Computer Science, or equivalent. Previous industrial and/or academic experience is desirable. The University welcomes and honors people of all races, genders, creeds, cultures, and sexual orientations, and values intellectual curiosity, pursuit of knowledge, and academic freedom and integrity. Interested candidates should apply through the departmental web site at http://www.eecs.utk.___ edu and submit a letter of application, a curriculum vitae, a statement of research and teaching interests, and provide contact information for three references. Consideration of applications will begin on January 1, 2008, and the position will remain open until filled. The University of Tennessee is an EEO/AA/Title VI/Title IX/Section 504/ ADA/ADEA institution in the provision of its education and employment programs and services. All qualified applicants will receive equal consideration for employment without regard to race, color, national origin, religion, sex, pregnancy, UNIVERSITY OF SOUTHERN CALIFORNIA Faculty Positions Ming Hsieh Department of Electrical Engineering The Ming Hsieh Department of Electrical Engineering (http://ee.usc.edu) of the USC Viterbi School of Engineering (http://viterbi.usc.edu) seeks outstanding faculty candidates for tenure track assistant professor, and tenured associate professor or professor positions. Areas of interest include but are not limited to: VLSI design; networking for ubiquitous computing and communication; wireless communications; information theory and coding; quantum information processing; control theory; bio and nano-signal processing; and cognitive signals and systems. Additional areas include: circuit design for analog/mixed signal processing; wireless data and power transmission for human-implantable systems; physical sensing and interfaces to biomolecular functions; nanoscale electronics; nanophotonics; applied nanoscience; and quantum engineering. Faculty members are expected to teach undergraduate and graduate courses, supervise undergraduate, graduate, and post-doctoral researchers, advise students, and develop a funded research program. Applicants must have a Ph.D. or the equivalent in electrical engineering or a related field and a strong research and publication record. Applications must include a letter clearly indicating area(s) of specialization, a detailed curriculum vitae, a one-page statement of current and future research directions and funding, and contact information for at least four professional references. The graduate program of the USC Viterbi School of Engineering is listed as seventh in the most recent US News and World Report rankings. The School’s 170 tenured and tenure track faculty includes 26 members of the National Academy of Engineering, seven winners of Presidential Early Career Awards, 45 Early Career Awards, four winners of the Shannon Award, and a co-winner of the 2002 Turing Award. USC faculty conduct research in leading-edge technologies with about $160 million in research expenditures annually. The Viterbi School is home to the Information Sciences Institute (ISI), two National Science Foundation-funded Engineering Research Centers (the Integrated Media Systems Center and the Biomimetic MicroElectronic Systems Center), the USC Stevens Institute for Technology Commercialization, and the Department of Homeland Security’s first center of excellence (CREATE). Applications and the names and email addresses of at least four professional references should be submitted electronically at http://ee.usc.edu/applications/faculty_position/faculty.php. USC is an Affirmative Action/Equal Opportunity Employer and encourages applications from women and members of underrepresented groups. 95 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page marital status, sexual orientation, age, physical or mental disability, or covered veteran status. HEWLETT – PACKARD COMPANY is accepting resumes for the following positions in San Diego, CA: Mechanical Design Engineer (Reference # SDRUL), Hardware Reliability Engineer (Reference # SDRRA), Electrical/Hardware Engineer (Reference # SDLPA). Please send resumes with reference number to HewlettPackard Company, 19483 Pruneridge Avenue, Mail Stop 4206, Cupertino, California 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. LEAD C++/SYBASE DEVELOPER - EGB Systems ,Stamford, CT - Develops complex derivative fixed income, workflow, pricing systems, Master in CS/Mathematics w/8 yrs exp in C++, sybase, SQL multithreading apply to careersusa@egb_________ systems.com. ________ HEWLETT – PACKARD COMPANY has an opportunity for the following position in Cupertino, California. Customer Oriented R&D Engineer. Resp. for providing adv. tech. support to the Support, Sales, Prof. Svc. organizations & customers. Reqs. education or knowledge of applications supported by system/network/web/database administrators; admin. & troubleshooting Windows desktop & server operating systems; Client/ Server systems & familiarity w/ HTTP tech.; UNIX admin: Sun Solaris, HP/UX, Red Hat Linux; Java, C, C++ programming; Network troubleshooting; & highavailability environments. Reqs. incl. Master’s degree or foreign equiv. in CS, Electrical & CE or related field of study. Send resume & refer to job #CUPLGI. Please send resumes with job number to Hewlett-Packard Company, 19483 Pruneridge Ave., MS 4206, Cupertino, CA 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. HEWLETT – PACKARD COMPANY has an opportunity for the following position in Cupertino, California. Systems/Software Engineer. Reqs. driver dev. skills, debugging, problem solving & troubleshooting skills in a Unix kernel environment; understanding internals of HPUX crash dump subsystem; implementation exp. w/ Storage Networking Industry Assoc. & Applic. Programming Interface. Reqs. incl. Master’s degree or foreign equiv. in CS, CE, Tech., Electrical/Electronic Eng. or related & 5 years of related exp. Send resume & refer to job #CUPKKA. Please send resumes with job number to Hewlett-Packard Company, 19483 Pruneridge Ave., MS 96 Computer 4206, Cupertino, CA 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. UNIVERSITY OF NORTH TEXAS, Computer Science and Computer Engineering Faculty Positions. The Department of Computer Science and Engineering at the University of North Texas invites applications and nominations for one faculty position at the level of Associate Professor to start in Fall 2008. Applicants are required to have expertise in Security and Information Assurance, Cyber Trust or a closely related area is required. The position requires an earned doctoral degree in Computer Science, Computer Engineering or a related field. Applicants must demonstrate an established record of quality teaching, funding history, graduate student supervision and national recognition. The CSE department offers a full complement of degrees in Computer Science and Computer Engineering. More information about the department can be found at http:// ____ www.cse.unt.edu/. Interested persons should send applications and nominations including a detailed curriculum vitae and have at least three letters of reference sent to: Faculty Search Committee, Department of Computer Science and Engineering, P.O. Box 311366, Denton, Texas, 76203, or electronically to: __ faculty_search@cse.unt.edu. ______________ The committee will begin its review of applications on December 1, 2007, and will continue to review applications once every month. The committee will accept applications until the position is filled, or the search is closed. The University of North Texas is an Equal Opportunity/Affirmative Action/ ADA employer, committed to diversity in its faculty and educational programs. HEWLETT – PACKARD COMPANY is accepting resumes for the following positions in Palo Alto, CA: Information Technology Developer/Engineer (Reference # PALBRA and #PALRAP), Pricing IT Functional Analyst/Designer (Reference # PALMMA), Electrical Hardware Engineer (Reference # PALDFA), Technical Analyst (Specialist) (Reference # PALDJA), Research Specialist (Reference # PALJJA). Please send resumes with reference number to Hewlett-Packard Company, 19483 Pruneridge Avenue, Mail Stop 4206, Cupertino, California 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. RESEARCH SCIENTIST IN DOCUMENT IMAGE ANALYSIS. FX Palo Alto Laboratory, Inc. (FXPAL) provides multimedia and collaboration technology research for Fuji Xerox Co., Ltd., a joint venture between Xerox Corporation of America A BEMaGS F and FujiFilm of Japan. We currently have an immediate opening for a Research Scientist with expertise in analysis of document images. Experience in document layout analysis, text analysis, graphics analysis, or in developing applications integrated with these types of analysis is desired. We are developing methods for extracting content from document images in English and Japanese and for using the extracted content in applications such as viewing and retrieval. The candidate should be interested in working on practical applications in a collaborative setting. Requires a Ph.D. in Computer Science or related field, strong development skills and excellent publication record. For more information about FXPAL, please see our site at www.fxpal. com. fxpalre___ To apply send resumes to: ____ sumes@fxpal.com and reference job code __________ CM/2. AUBURN UNIVERSITY, Department of Computer Science and Software Engineering, Assistant/Associate Professor. The Department of Computer Science and Software Engineering (CSSE) invites applications for one tenuretrack faculty position at the Assistant or Associate Professor level to begin Fall 2008. We encourage candidates from all areas of computer science and software engineering to apply. The following are preferred research areas: artificial intelligence, simulation, information assurance and security, database systems, theory, programming languages, and software engineering. The candidate selected for this position must be able to meet eligibility requirements to work in the United States at the time appointment is scheduled to begin and continue working legally for the proposed term of employment and be able to communicate effectively in English. Applicants should submit a current curriculum vita, research vision, teaching philosophy, and the names and addresses of three references to: Dr. Kai H. Chang, Chair, Computer Science and Software Engineering, Auburn University, AL 36849-5347. kchang@eng.auburn.edu _______________ (with copy to _____________ mccorba@auburn.edu), 334-844-6300 (Voice). The applicant review process will begin January 15, 2008. Detailed announcement of this position can be found at: http://www.eng.auburn.edu/ csse/. ___ Auburn University is an Affirmative Action/Equal Opportunity Employer. Women and minorities are encouraged to apply. UNIVERSITY AT BUFFALO, THE STATE UNIVERSITY OF NEW YORK, Faculty Positions in Computer Science and Engineering. Celebrating its 40th anniversary this year, the CSE Department solicits applications from excellent candidates in pervasive computing and high performance computing for openings at Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page the assistant professor level. The CSE department has outstanding faculty and is affiliated with successful centers devoted to biometrics, bioinformatics, biomedical computing, cognitive science, document analysis and recognition, and computer security. Candidates are expected to have a Ph.D. in Computer Science/Engineering or related field by August 2008, with an excellent publication record and potential for developing a strong funded research program. All applications should be submitted by January 15, 2008 electronically via recruit.cse.buffalo.edu. A cover letter, cur_____________ riculum vitae, and names and email addresses of at least three references are required. The University at Buffalo is an Equal Opportunity Employer/Recruiter. NOKIA SIEMENS NETWORKS US LLC has the following exp/degree positions in the following locations. Travel to unanticipated U.S. worksites may be required. EOE. Irving, Texas. •Solutions Manager: Provide sales and marketing support and solutions to customers; identify and qualify business opportunities and strategies; and define and document technical solutions. ID# NSN-TX-SM. •Systems Engineer: Test and troubleshoot network elements with new SW rollouts and customer trials, or design test cases and create plan for Network Interoperability Testing. ID# NSN-TX-SE. Chicago, Illinois. •Services Lead: Perform GSM engineering operations including network build-out, deployment; operations and management, services, scoping, procurement, and customer telecom management. ID# NSN-IL-SL. Herndon, Virginia. •Service Solution Manager: Manage service products portfolios; provide solution consulting; anticipate customer needs; identify solutions; and anticipate internal/external business issues and developments. ID# NSN-VA-SSM. Mail resume to: NSN Recruiter, Nokia Siemens Networks, 6000 Connection Dr., 4E-338, Irving, TX 75039. Must reference ID #. BOSTON UNIVERSITY. The Department of Electrical and Computer Engineering (ECE) at Boston University anticipates openings for faculty positions at all ranks in all areas of electrical and computer engineering. Areas of particular interest in computer engineering are computer systems, embedded and reconfigurable systems, distributed systems, trusted computing, design automation, VLSI, computer networks, software engineering, and related areas. The ECE Department is part of a rapidly developing and innovative College of Engineering. Excel- A BEMaGS F lent opportunities exist for collaboration with colleagues in outstanding research centers at Boston University, at other universities/colleges, and with industry throughout the Boston area. The Department has 44 faculty members, 200 graduate students and 250 BS majors. For additional information, please visit http://www.bu.edu/ece/. In addition to a relevant, earned PhD, qualified candidates will have a demonstrable ability to teach effectively, develop funded research programs in their area of expertise, and contribute to the tradition of excellence in research that is characteristic of the ECE department. Applicants should send their curriculum vita with a statement of teaching and research plans to: Professor David Castañón, Chair ad interim, Department of Electrical and Computer Engineering, Boston University, 8 Saint Mary’s Street, Boston, MA 02215. Boston University is an Equal Opportunity/Affirmative Action Employer. HEWLETT – PACKARD COMPANY has an opportunity for the following position in Richardson, Texas. Technical Consultant. Resp. for providing expert consulting to external company customers in the SAP Business Intelligence (BI) area. Reqs. in-depth business knowledge, expert SAP BI knowledge & travel to TECHNOLOGY IN SUPPORT OF NATIONAL SECURITY MIT Lincoln Laboratory is a premier employer applying science and advanced technology to critical, real-world problems of national interest. Computer Scientists Computer Scientists are needed to conduct applied research on infrastructures for large distributed sensing, communications, and decision making systems. Candidates should have experience and interest in several of the following areas: distributed systems and applications, autonomous intelligent agents, knowledge representation and ontology’s, knowledge-based systems, databases, semantic web and semantic web services, software and systems architecture, service oriented architecture (SOA), and user interfaces. _______________ A recent PhD or an MS with at least 2 years experience in Computer Science or related field is desired. ___________ Please apply online at: __________ http://www.ll.mit.edu/careers/careers.html. Lincoln Laboratory is an Equal Opportunity Employer. M/F/D/V - U.S. Citizenship required. 97 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page worksites throughout the U.S. Reqs. incl. Bachelor’s degree or foreign equiv. in CS, Eng., Electrical/Electronic Eng. or related & 5 years of related exp. Send resume & refer to job #RICJVA. Please send resumes with job number to Hewlett-Packard Company, 19483 Pruneridge Ave., MS 4206, Cupertino, CA 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. UNIVERSITY OF WASHINGTON, Computer Science & Engineering, Tenure-Track, Research, and Teaching Faculty. The University of Washington's Department of Computer Science & Engineering has one or more open positions in a wide variety of technical areas in both Computer Science and Computer Engineering, and at all professional levels. A moderate teaching and service load allows time for quality research and close involvement with students. Our recent move into the Paul G. Allen Center for Computer Science & Engineering expands opportunities for new projects and initiatives. The Seattle area is particularly attractive given the presence of significant industrial research laboratories as well as a vibrant technology-driven entrepreneurial community that further enhance the intellectual atmosphere. Information about the department can be found on the web at http://www.cs.washington.edu. We welcome applicants in all research areas in Computer Science and Computer Engineering including both core and inter-disciplinary areas. We expect candidates to have a strong commitment both to research and to teaching. The department is primarily seeking individuals at the tenure-track Assistant Professor rank; however, under unusual circumstances and commensurate with the qualifications of the individual, appointments may be made at the rank of Associate Professor or Professor. We also seek non-tenured research faculty at Assistant, Associate and Professor levels, and full-time annual Lecturers and Sr. Lecturers. Applicants for both tenure-track and research positions must have earned a PhD. by the date of appointment; those applying for lecturer positions must have earned at least a Master’s degree. Please apply online at <http://www.cs.washington.edu/news/jo bs.html> with a letter of application, a _____ complete curriculum vitae, statement of research and teaching interests, and the names of four references. Applications received by February 29, 2008 will be given priority consideration. The University of Washington was awarded an Alfred P. Sloan Award for Faculty Career Flexibility in 2006. In addition, the University of Washington is a recipient of a National Science Foundation ADVANCE Institutional Transformation Award to increase the participation of women in academic science and engineering careers. The Uni- 98 Computer versity of Washington is an affirmative action, equal opportunity employer. We are building a culturally diverse faculty and encourage applications from women, minorities, individuals with disabilities and covered veterans. RUTGERS UNIVERSITY, Tenure Track Faculty Position in Computational Biology or Biomedical Informatics. The Department of Computer Science and the BioMaPS Institute for Quantitative Biology at Rutgers University invite applications for a tenure track faculty position at the junior or senior level in the Department of Computer Science. Candidates should have a strong background and experience in computational biology or biomedical informatics, including but not limited to: structural and functional genomics and proteomics, biological networks, evolutionary and systems biology, computational modeling, machine learning and applications, large scale systems data analysis, and informatics. They should be prepared to work on interdisciplinary projects making substantive Computer Science contributions. Applicants should submit a cover letter, curriculum vitae, research summary and statement of future research goals, together with a statement of teaching experience and interests, and arrange for four letters of recommendation to be sent on their behalf. Materials should be sent as PDF files to: Chair, Hiring Committee DCS-BioMaPS, Rutgers University, Department of Computer Science, Hill Center, Busch Campus, Piscataway, NJ 08855 (email: _______________ hiringbio@cs.rutgers.edu ). For more information on the Department of Computer Science, see http://www.cs.rutgers.edu, and for the BioMaPS Institute, see __________ http://www.biomaps.rutgers.edu. To ensure proper con__________ sideration, applications should arrive by February 1, 2008. Rutgers University is an Affirmative Action/Equal Opportunity Employer. Women and minority candidates are especially encouraged to apply. .Net DEVELOPER – EGB Systems ,Stamford, CT - Designs, develops & tests .Net applications in C#, C++, VB, asp, XML, COM, DCOM, DTD, XSL, XSLT, CSS, Oracle, MS/SQL server & .Net. Bachelor in Science/ Eng w/ 5+ yrs exp in .Net tech apply to careersusa@egbsystems.com. _________________ INFORMATION SYSTEMS ANALYST. Analyze & design info. mgmt. sys. using LAN. Dvlp. utilizing MS SQL. Troubleshoot, recover & maintain OS on 2000NT Server & networks. Analyze auto accessories data including monthly sales, financial data and performance reports to President. Req: Master's in Info. Sys. Mgmt., CIS., or Bus. Admin w/courses in Computers. 40 hr/wk. Job/Interview Site: City of Industry, CA. Send Resume to: TC A BEMaGS F Spoilers, Inc. @ 228 Sunset Ave, City of Industry, CA 91744. COMPUTER WEB DESIGN & DEVELOPMENT COMPANY (San Francisco, CA) seeking Web Designer/ Developer. Reqs MS in Web D&D +6 mos exp in PHP, Perl, AJAX, JavaScript, ActionScript, XHTML, XML, Shell. S/ware: SSH, Flash, Dreamweaver, Photoshop, CorelDraw, WebPosition, Visio, MS Office. Apply to Roman Bogomazov, President, Papillon Lab Corp, 1324 Clement St, San Francisco, CA 94118. HEWLETT – PACKARD COMPANY has an opportunity for the following position in Cupertino, CA. Software Design Engineer. Reqs. exp with SW Dvlpmt in a Unix envir, C prgrm lang. Script lang e.g. PERL, Unix Shell prgrm. Unix kernel internals. Device driver dvlpmt. Must have completed at least 2 full life cycle projects. Mass storage e.g. SCSI & Fiber Channel. Reqs. incl. Bachelor’s degree or foreign equiv. in CS, CE, EE or related & 5 yrs of related exp. Employer will accept a combination of foreign degrees and/or diplomas and/or Certifs determined to be equiv to a U.S. bachelor’s degree. Send resume & refer to job #CUPTMA. Please send resumes with job number to Hewlett-Packard Company, 19483 Pruneridge Ave., MS 4206, Cupertino, CA 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. SR. TECH. CONSULTANT, Ascendant Technology, Austin, TX. Req.: Master's (or equiv.) in Comp. Sci., IT, Engineering, Math., rel. or foreign equiv. Resume only attn.: C. Jones (#06-1036) 10215 161st Pl. NE, Redmond, WA 98052. THE UNIVERSITY OF KANSAS, Faculty Position in Bioinformatics/ Computational Biology. The Center for Bioinformatics (www.bioinformatics. ____ and the Department of Electrical ku.edu) Engineering and Computer Science (http://www.eecs.ku.edu) at The University of Kansas invite applications for a tenure-track assistant professor position expected to begin August 18, 2008. The Bioinformatics initiative is part of a major expansion in Life Sciences and complements existing strengths in information technology, structural biology, computational chemistry, biophysics, proteomics, developmental and molecular genetics, and drug design. Duties: to establish and maintain an externally-funded research program, to participate in teaching, and to provide service. Required Qualifications: Ph.D. in a discipline related to Bioinformatics or Computer Science expected by start date of appointment; potential for excellence in research in Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Bioinformatics; and commitment to teaching bioinformatics and computer science courses; strong record of research accomplishments in at least one of the following areas: biomolecular networks modeling, bioinformatics databases, and computational modeling of biomolecular systems. For the full position announcement, refer to: http://www2. ku.edu/~clas/employment/FY09_Index. _______________________ htm and click the Bioinformatics/EECS __ download button. E-mail application as a single file, including CV, letter of application, statement of past and future research and teaching interests and philosophy to: __________ asawyer@ku.edu. Have at least three letters of reference sent separately to: Dr. Robert F. Weaver, Professor and Associate Dean, College of Liberal Arts and Sciences, c/o Anne Sawyer, The University of Kansas, 1450 Jayhawk Blvd, 200 Strong Hall, Lawrence, KS 660457535. Initial review of applications begins January 15, 2008 and will continue until the position is filled. EO/AA Employer. HEWLETT – PACKARD COMPANY has an opportunity for the following position in Palo Alto, California, and various unanticipated sites throughout the United States. Software System Engineer. Reqs. Master’s in CS, CE, Engineering or related and 3 yrs related exp. Reqs. Web Application program/design, J2EE, XML, SOAP, C++. Send resume referencing #PALPBR. Please send resumes with reference number to Hewlett-Packard Company, 19483 Pruneridge Ave., MS 4206, Cupertino, CA 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. GEORGETOWN UNIVERSITY, Senior Faculty Position and Chair of Department, Department of Computer Science. The Department of Computer Science seeks a dynamic scholar/teacher for a senior faculty position within the department. It is expected that within a short time of coming to Georgetown, this new faculty member will assume the duties and responsibilities of department chair. For more information, please visit our website at ____ http:// www.cs.georgetown.edu/. DATABASE ADMINISTRATOR. Analyze, dsgn, dvlp, test & implmt data warehouses, web based applics & applic systems & s/w using 9I, 10 G w/RAC, Data Guard, Physical standby DB, Toad, Sql loader, exp/imp, Vertias Netbackup, SQL Dvlpr, OEM, PL/SQL, RMAN & UNIXSolaris/AIX/Linux. Ability to perform & troubleshoot advance Oracle DBA functions such as backups, recovery, tuning, monitoring, scripting etc. Provide tech assistance in connection w/system main- tenance; prep technical documentation for user ref; conduct training sessions for end users. Reqs MS in Comp Sci or related. Mail resumes to MCS Technology Solutions LLC, 200 Centennial Ave, Ste 200, Piscataway, NJ 08854. DATABASE ADMINISTRATOR sought by Importers/Wholesalers. Respond by resume only to: J. Wong, ITC Intertrade, Inc., 1505 Sawyers St., Suite B. Houston, TX 77007. HEWLETT – PACKARD COMPANY is accepting resumes for the following positions in Houston, TX: Business Systems Analyst (Reference # HOURPU), Team Leader/Architect (Reference # HOUEAL), eCommerce IT Business Analyst (Reference # HOUBJO). Please send resumes with reference number to HewlettPackard Company, 19483 Pruneridge Avenue, Mail Stop 4206, Cupertino, California 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. SQL DEVELOPER - EGB Systems, Stamford, CT - Develops SQL database applications & stored procedures. Knowledge of , Java, J2EE, JSP, JSF Portal, Java Script, HTML, XML, TOAD, WIN2000/XP, Crystal Report 9.0 required. Master in CS w/5+ yrs exp in SQL server apply to careersusa@egbsystems.com. _________________ RUTGERS UNIVERSITY, Tenure Track Faculty Position in Computational Biomedicine, Imaging and Modeling. The Rutgers University Department of Computer Science and the Center for Computational Biomedicine, Imaging and Modeling (CBIM) seeks applicants in computer graphics and related areas, for a tenure-track faculty position starting September 2008. We're particularly interested in synergy with CBIM and thus we’re excited about receiving applications primarily in all areas of computer graphics, as well as related areas such as visualization, computer vision, machine learning, and human-computer interaction. Rutgers University offers an exciting and multidisciplinary research environment and encourages collaborations between Computer Science and other disciplines. Applicants should have earned or anticipate a Ph.D. in Computer Science or a closely related field, should show evidence of exceptional research promise, potential for developing an externally funded research program, and commitment to quality advising and teaching at the graduate and undergraduate levels. Applicants should send their curriculum vitae, a research statement addressing both past work and future plans, a teaching statement and arrange for four letters of recommenda- A BEMaGS F tion to be sent on their behalf to __ hiring@cs.rutgers.edu. If electronic submis___________ sion is not possible, hard copies of the application materials may be sent to: Professor Dimitris Metaxas, Hiring Chair, Computer Science Department, Rutgers University, 110 Frelinghuysen Road, Piscataway, NJ 08854, Applications should be received by February 15, 2008, for full consideration. Rutgers University is an Affirmative Action/Equal Opportunity Employer. Women and minority candidates are especially encouraged to apply. TENNESSEE TECHNOLOGICAL UNIVERSITY. The Department of Electrical and Computer Engineering (ECE) at Tennessee Technological University invites applications for a tenure-track position at the Associate or Assistant Professor level in Computer Engineering beginning in August 2008. This position includes an initial appointment as a Stonecipher Faculty Fellow in Computer Engineering. Candidates for this position must have an earned Ph.D. degree in computer engineering or closely related areas with expertise in computer engineering areas such as computer networks, parallel and distributed systems, sensor networks and/or computer security. Screening of applications will begin February 1, 2008 and continue until the position is filled. See http://www.tntech.edu/ece/jobs. html for details. AA/EEO. ___ CDI CORP seeks IT Architect/Specialist for Memphis, TN to provide varied set of networking systems services in support of complex networks/systems environments. Req BS, 1 yr technical work exp and Cisco CCNA or CCNP cert. Mail resume ATTN: Generalbox 1960 Research Drive, Suite 200 Troy MI 48083 EOE. HEWLETT - PACKARD COMPANY is accepting resumes for the following position in Houston, TX: IT Analyst (Business Systems Analyst) (Reference # HOUVKU). Please send resumes with reference number to Hewlett-Packard Company, 19483 Pruneridge Avenue, Mail Stop 4206, Cupertino, California 95014. No phone calls please. Must be legally authorized to work in the U.S. without sponsorship. EOE. PROGRAMMER ANALYST: Jobsite: Sacramento CA. MS Electrical Engineering or Computer Engineering required. Send resume: Logic House Ltd. 78-670 Highway 111 #218, La Quinta, CA 92253. COMPUTER & INFO SYSTEMS MANAGER sought by Chicago-based co. Create in house prgm for employee use: for referral, rehab, payroll & frequency, mktg, 99 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page mailing, billing, supervisor nurse & administrator depts, to benefit co. Min reqmt 1 yr related work exp & BS in Comp Info System. Send resumes to Angela, Healing Hands Home Care, Inc, 6730 W. Higgins Ave, Chicago, IL 60656. SR. LEAD CONSULTANT: Lead implemntation for supply chain, dist., create financial modules of Oracle apps. Implemnt; plan project, estimate, track, reprt; resolve custmer issues; develop fnctional specs; maintain opratnal documntatn. Train users, post-prod supprt, design custom flows/specs for extensns. BS/BA Comp Sci, Commerce, Bus. + 2 yrs exp as System Anal, Prog. Anal, or any related in IT field, exp must incl. requirements gathering, analysis, process design, app config, systm test, industry exp and consult exp w/ Oracle dist. apps., use of Oracle distrib. suite. Will accept any suitable comb of edu, training, exp. Arabella Lam, RCM Technologies, 1055 W. 7th Street, Suite 1820, Los Angeles, CA 90017. Ref. Code RCMPL1. Include ref. code. GEORGE MASON UNIVERSITY, Department of Computer Science, Volgenau School of Information Technology and Engineering, Faculty Positions in Bioengineering. The Volgenau School of Information Technology and Engineering at George Mason University is building a program in bioengineering, including computational approaches to biology. As part of this multidisciplinary initiative, tenure track openings are available at the Assistant, Associate and Full Professor levels in the School. Background, experience, and interest will determine the departmental affiliation of successful applicants. Within this initiative, the Department of Computer Science is seeking faculty members who can establish strong research and teaching programs in the area of computational biology, bioinformatics, and biometrics. Minimum qualifications include a Ph.D. in Computer Science, Bioinformatics, or a closely related field, demonstrated potential for excellence and productivity in research applying computational approaches to address fundamental questions in biology or medicine, and a commitment to high quality teaching. Candidates for a senior position must have a strong record of external research support. The School has more than 100 full-time faculty members, with over 40 in the CS Department. The research interests of the CS department includes artificial intelligence, algorithms, computational biology and bioinformatics, computer graphics, computer vision, databases, data mining, image processing, security, knowledge engineering, parallel and distributed systems, performance evaluation, real-time systems, robotics, software engineering, visualization, and wireless and mobile computing. 100 Computer The department has several collaborative research and teaching activities in computational biology and bioinformatics with other Mason units. For more information, visit our Web site: http://cs. gmu.edu/. For full consideration please ______ submit application and application materials on-line at http://jobs.gmu.edu (position number F9086z). To apply online you will need a statement of professional goals including your perspective on teaching and research, a complete CV with publications, and the names of four references. The review of applications will begin immediately and will continue until the positions are filled. George Mason University is a growing, innovative, entrepreneurial institution with national distinction in several academic fields. Enrollment is 30,000, with students studying in over 100 degree programs on four campuses in the greater Washington, DC area. Potential interactions with government agencies, industry, medical institutions, and other universities abound. GMU is an equal opportunity/affirmative action employer that encourages diversity. IOWA STATE UNIVERSITY OF SCIENCE AND TECHNOLOGY, College of Liberal Arts and Sciences, Announcement of Opening for Tenure-track Position. The Department of Computer Science at Iowa State University is seeking outstanding candidates to fill a tenure-track position, to commence in August, 2008. We are especially interested in applicants at the assistant professor level in Programming Languages and/or Software Engineering. Successful candidates will have demonstrated potential for outstanding research and instruction in computer science. A Ph.D. or equivalent in Computer Science or a closely related field is required. Our department currently consists of 27 fulltime tenure-track faculty members. We offer B.S., M.S., and Ph.D. degrees in Computer Science and participate in new B.S. degrees in Software Engineering and in Bioinformatics and Computational Biology. We also participate in interdepartmental graduate programs in Bioinformatics and Computational Biology, Human-Computer Interactions, and Information Assurance. We have about 330 B.S. students, 60 M.S. students, and 110 Ph.D. students. Almost all graduate students are supported by research or teaching assistantships. We have strong research and educational programs in Algorithms and Complexity, Artificial Intelligence, Bioinformatics and Computational Biology, Databases, Data Mining, Information Assurance, Programming Languages, Multimedia Systems, Operating Systems and Networks, Robotics, and Software Engineering. Our department has over $6.5 million in active research grants. With the above interdisciplinary activities included, we con- A BEMaGS F tribute to active research and training grants totaling approximately $20 million. A dynamic faculty, a moderate teaching load (typically 3 courses per year with one course reduction for active researchers and possible further reductions for junior faculty), a strong graduate program, and a well-funded research program provide an excellent academic environment. In addition, cutting-edge research and education are nurtured through interdisciplinary interactions facilitated by the Laurence H. Baker Center for Bioinformatics and Biological Statistics, the Center for Computational Intelligence, Learning and Discovery, the Center for Integrative Animal Genomics, the Cyber Innovation Institute, the Information Assurance Center, the Department of Energy's Ames Laboratory, and the Virtual Reality Application Center. Iowa State University is a major land-grant university located in Ames, Iowa. It is a pleasant, small, cosmopolitan city with a population of over 50,000 (including about 27,000 students), a vibrant cultural scene, an excellent medical clinic, and a secondary school system that ranks among the best in the United States. Ames is frequently ranked among the best communities to live in North America: 20th nationally among best places to live (2002), 3rd nationally in terms of highly educated workforce for knowledge-based industry (2005), 12th nationally for its public schools (2006). Applicants should send a curriculum vita, including teaching and research statements and the names and addresses of at least three references, to: Chair of Search Committee, Department of Computer Science, Iowa State University, Ames, Iowa 50011-1041; Fax: 515-294-0258; Tel: 515-294-4377; E-mail: _____ facultysearch@cs.iastate.edu; Web: _______ www.cs.ias_____________ tate.edu. Review of applications will begin _____ on December 1, 2007 and will continue until the position is filled. Iowa State University is an equal opportunity employer. Women and members of underrepresented minorities are strongly encouraged to apply. For more information, please visit us at http://www.cs.iastate. edu. __ FUTURE TECHNOLOGY ASSOCIATES in NYC seeks Consultant to design, develop, deploy & maintain BizTalk, SQL & EDI solutions for in- & out-bound documents; Req BS/BA and 1 yr BizTalk Architecture & development, 1 yr w/ .NET, SQL, XML/XSLT & MQ Series BizTalk client configuration & 6 mos. w/ EDI & Microsoft MOM. Email resume to tsevint@schools.nyc.gov. ______________ SYSTEM SOFTWARE ENGINEER: MS in Electrical or Computer Engineering req. Send resume: Sundance Digital Signal Processing Inc., 4790 Caughlin Pkwy #233 , Reno, NV 89509. Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page TEXAS TECH UNIVERSITY: The Department of Computer Science invites applications for one or more Abilene faculty positions at any rank starting Fall 2008. Preferences are given to candidates with background in software architecture, formal methods, verification, validation, and security; web-based, trusted and intelligent software development. Other areas will be considered for excellent candidates. Applicants must have a Ph.D. degree in computer science or a closely related field. Successful candidates must have demonstrated achievements or potentials for excellence in research and teaching. The Abilene facility offers only graduate programs at the masters and the doctoral levels to students in Abilene and other cities via ITV and distance learning. The department has received funding support from the Abilene foundation to provide research assistantships in Abilene. We offer competitive salaries, a friendly and cooperative environment, and excellent research facilities. Review will begin in January 2008 and continue until the positions are filled. A letter of application, curriculum vitae, brief research and teaching goals, and three letters of reference should be submitted electronically at http://jobs.texastech. edu. Please use Requisition number ___ 75218. Additional information is available at http://www.cs.ttu.edu. Texas Tech University is an equal opportunity/affirmative action employer and actively seeks applications from minority and female applicants. GEORGE MASON UNIVERSITY, Department of Computer Science. The Department of Computer Science at George Mason University invites applications for a tenure-track faculty position at the rank of Assistant Professor beginning Fall 2008. We are seeking a faculty member who can establish strong research and teaching programs in the area of computer game development. Applicants must have a research focus in an area in computer games technology — for example, in artificial intelligence, computer graphics, real-time animation, simulation and modeling, distributed and multi-agent systems, or software engineering, as applied to computer games. Minimum qualifications include a Ph.D. in Computer Science or a related field, demonstrated potential for excellence and productivity in research, and a commitment to high quality teaching. The department currently offers a graduate certificate in Computer Games Technology, and is adding a concentration in Computer Game Design to its undergraduate program. The Computer Game Design concentration is being developed in close collaboration with faculty in the College of Visual and Performing Arts at Mason. For more information on these and other programs offered by the department, visit our Web site: http://cs.gmu.edu/. The department has over 40 faculty members with wide-ranging research interests including artificial intelligence, algorithms, computer graphics, computer vision, databases, data mining, security, human computer interaction, parallel and distributed systems, real-time systems, robotics, software engineering, and wireless and mobile computing. George Mason University is located in Fairfax, Virginia, a suburb of Washington, DC, and home to one of the highest concentrations of high-tech firms in the nation. There are excellent opportunities for interaction with government agencies and industry, including many game and “serious game” development companies. In particular, the Washington DC region is fast becoming a hub for the serious games industry. Fairfax is consistently rated as being among the best places to live in the country, and has an outstanding local public school system. For full consideration please submit application and application materials on-line at http://jobs.gmu.edu (position number F9084z). To apply, you will need a statement of professional goals including your perspective on teaching and research, a complete C.V. with publications, and the names of four references. The review of applications will begin immediately and will continue until the position is filled. GMU is an equal opportunity/affirmative action employer. Women and minorities are strongly encouraged to apply. CAL POLY, SAN LUIS OBISPO, Computer Engineering. The Computer Science Department and Computer Engineering Program, at Cal Poly, San Luis Obispo, invite applications for a full-time, academic year, tenure track Computer Engineering faculty position beginning September 8, 2008. Rank and salary is commensurate with qualifications and experience. Duties include teaching core undergraduate courses, and upper-division and master's level courses in a specialty area; performing research in a mainstream area of computer engineering; and service to the department, the university, and the community. Applicants from all mainstream areas of computer engineering are encouraged to apply. A doctorate in Computer Engineering, Computer Science, Electrical Engineering, or a closely related field is required. Candidates in the areas of architecture and parallel computing are especially encouraged to apply. Candidates must have a strong commitment to teaching excellence and laboratory-based instruction; dedication to continued professional development and scholarship; and a broad-based knowledge of computer engineering. Demonstrated ability in written and oral use of the English language is required. Computer Engineering is a joint program between the Departments of Computer Science and Electrical Engineering. Cal Poly offers Bachelor's A BEMaGS F Degrees in Computer Engineering, Computer Science, Software Engineering and Electrical Engineering, and Master's Degrees in Computer Science and Electrical Engineering. Cal Poly emphasizes "learn by doing" which involves extensive lab work and projects in support of theoretical knowledge. The available computing facilities for instructional and faculty support are modern and extensive. To apply, please visit ____ WWW. CALPOLYJOBS.ORG to complete a ____________ required online faculty application, and apply to Requisition #101387. For full consideration, candidates are required to attach to their online application: (1) resume, (2) cover letter, (3) candidate's statement of goals and plans for teaching and research. Upon request, candidates selected as finalists will be required to submit three letters of reference and official transcripts for final consideration. Review of applications will begin January 7, 2008; applications received after that date may be considered. Questions can be emailed to: _________________ cpe-recruit@csc.calpoly.edu. Please include requisition #101387 in all correspondence. For further information about the department and its programs, see www.csc.calpoly.edu and www.cpe. calpoly.edu. Cal Poly is strongly commit_______ ted to achieving excellence through cultural diversity. The university actively encourages applications and nominations of all qualified individuals. EEO. Online Advertising Are you recruiting for a computer scientist or engineer? Submission Details: Rates are $160.00 for 30 days with print ad in Computer magazine. Send copy to: Marian Anderson IEEE Computer Society 10662 Los Vaqueros Circle Los Alamitos, California 90720-1314; phone: + 1 714.821.8380 fax: +1 714.821.4010; email: ______________ manderson@computer.org. http://computer.org 101 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F BOOKSHELF aturing Usability: Quality in Software, Interaction and Value, Effie Law, Ebba Hvannberg, and Gilbert Cockton, eds. This book’s essays supply an understanding of how current research and practice contribute to improving quality from the perspectives of software features, interaction experiences, and achieved value. Divided into three parts, the book analyzes how using development tools can enhance system usability and how methods and models can be integrated into the development process to produce effective user interfaces. The essays address theoretical frameworks on the nature of interactions, techniques, and metrics for evaluating interaction quality, and the transfer of concepts and methods from research to practice. They also assess the impact a system has in the real world, focusing on increasing the value of usability practice for software development and on increasing value for users. Springer; www.springer.com; 9781-84628-940-8; 429 pp. M overnance and Information Technology: From Electronic Government to Information Government, Viktor Mayer-Schönberger and David Lazer, eds. Developments in information and communication technology and networked computing over the past two decades have given rise to the notion of electronic government, which most commonly refers to the delivery of public services over the Internet. This volume argues for a shift from electronic government’s narrow focus on technology and transactions to the broader perspective of information government—the information flows within the public sector, between the public sector and citizens, and among citizens—as a way to understand the changing nature of governing and governance in an information society. The essays describe the interplay between recent technological developments and evolving information flows and the implications of different information flows for efficiency, political mobilization, and democratic accountability. G 102 MIT Press; ____________ mitpress.mit.edu; 0-26263349-3; 352 pp. haracter Recognition Systems: A Guide for Students and Practitioners, Mohamed Cheriet, Nawwaf Kharma, Cheng-Lin Liu, and Ching Y. Suen. This book provides practitioners and students with the fundamental principles and state-of-the-art computational methods for reading printed texts and handwritten materials. The authors present information analogous to the stages of a computer recognition system, helping readers master the theory and latest methodologies used in character recognition. Each chapter contains major steps and tricks to handle the tasks described. Researchers and graduate students in computer science and engineering might find this book useful for designing a concrete system in OCR technology, while practitioners might find it a valuable resource for the latest advances and modern technologies not covered elsewhere in a single book. Wiley-Interscience; www.wiley.com; 978-0-471-41570-1; 360 pp. C igital Convergence—Libraries of the Future, Rae Earnshaw and John Vince, eds. The convergence of IT, telecommunications, and media is revolutionizing how information is collected, stored, and accessed. Digital information preserves content accuracy in a way other systems do not. Highbandwidth transmission from one place to another on the planet is now possible. Ubiquitous and globally accessible, information can be held and accessed just as easily on a global network as on a local personal computer or in a local library. Devices are increasingly intelligent and network-ready, while user D interfaces become more adaptable and flexible. Digital intelligence is becoming seamless and invisible, enabling a greater focus on content and the user’s interaction with it. This revolution affects the development and organization of information and artifact repositories such as libraries, museums, and exhibitions, while also changing how physical and digital aspects are mediated. Digital convergence is bringing about changes that are substantial and likely to be long-lasting. This volume presents key aspects in the areas of technology and information sciences, from leading international experts. Springer; www.springer.com; 9781-84628-902-6; 416 pp. jax in Action, Dave Crane and Eric Pascarello with Darren James. Web users are tiring of the traditional Web experience. They get frustrated losing their scroll position, become annoyed waiting for refresh, and struggle to reorient themselves on every new page. With asynchronous JavaScript and XML, known as Ajax, users can enjoy a better experience. Ajax is a new way of thinking that can result in a flowing and intuitive interaction with the user. This book explains how to distribute the application between the client and server—by using a nested MVC design, for example—while retaining the system’s integrity. It also shows how to make the application flexible and maintainable using good, structured design to help avoid problems like browser incompatibilities. Above all, this book showcases the many advantages gained by placing much of the processing in the browser. Developers with prior Web technologies experience might find this book especially useful. Manning; www.manning.com/ crane; 1-932394-61-3; 680 pp. A Send book announcements to newbooks@ computer.org. _________________ Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F SECURITY Natural-Language Processing for Intrusion Detection Allen Stone, Jacob & Sundstrom EBIDS-SENLP, a system using natural-language processing, shows promise in defending against social-engineering attacks. I ntrusion-detection systems seek to electronically identify malicious traffic as it enters a defended network. Traditionally, IDSs have used electronic signatures or traffic analysis to identify attack traffic. Spam filtering keys on specific electronic segments of an e-mail and filters out those identified as possibly malicious. Social engineering, a unique type of attack traffic, attempts to compromise a network or system’s security metrics by exploiting the human end user through natural language, based on common psychological flaws and deception. These attacks have been difficult to defend against in the past with IDSs because natural language is highly variable. An attacker could easily reword the text, while preserving the attack’s same underlying conceptual meaning. Since this type of attack exploits the end user, personnel resistance training has been the most effective means of defense, but all training wears off over time, and even properly trained personnel can make mistakes. Natural-language processing teaches computers the semantic meaning of natural-language text. Thus, an NLP system reads plain English (among other languages) and categorizes what it’s seen in terms of conceptual themes and ontological concepts. The University of Maryland, Baltimore County, is home to an advanced NLP system, OntoSem. until I could modify and streamline it for this unintended purpose. Defining rules for EBIDS EBIDS, which I developed as part of my master’s thesis at UMBC, simply parses e-mail text and strips out the body. It sends the message to a detection mechanism, then uses a threshold system to decide if the e-mail is likely malicious. The real value in EBIDS, as with most IDSs, is in the definition of its rules. EBIDS rules are based on the concept of deception, rather than literal string matches, although it can incorporate literal strings where it would be best to do so. The concepts of deception are more or less an equivalence set of literal strings, using OntoSem’s categorization to build these sets. The language for rules allows for detailed and complex signatures. Each rule comprises some number of signatures in the following format: • a comment field; • either “SEM” for semantic analysis, “REF” for referential analysis, or “LIT” for literal string; • the contextual concept of deception that the signature looks for; and • a weighting to the rule. SOCIAL-ENGINEERING ATTACKS Social-engineering defense is flawed because it depends on humans. Effectively training the system—rather than the end user—would eliminate that flaw. The E-mail-Based Intrusion Detection System to find Social Engineering using Natural-Language Processing (EBIDS-SENLP) does this by reading plain-text e-mails and sending the e-mail’s natural-language body to OntoSem. However, OntoSem needs pretraining to look for certain segments of language as concepts of deception to meaningfully analyze the raw text and tell EBIDS what types of deception the e-mail could contain. OntoSem proved too time-inefficient for the project’s testing, so I used an equivalence set to work around this The F.F. Poirot project attempts to find phishing e-mails sent to financial organizations using a similar ontology, but the structure of their rules is rigidly defined, and the narrow focus on financial groups is difficult to expand. EBIDS isn’t restricted to financial fraud, has an easily updated ontology (OntoSem), and doesn’t require as rigid a rule definition. Figure 1 shows an example of an EBIDS detection rule. Rules for testing The specific rules for testing EBIDS focus primarily on phishing, because phishing e-mails entirely comprised the “known bad” corpus. However, my intent is to expand the system to other types of social engineering. 103 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F SECURITY cards, congratulatory language, a corporation name, and an information request. Triggers Kevin Mitnick describes the psychological triggers these attacks typically use in The Art of Deception: Controlling the Human Element of Security (Wiley, 2002). These triggers include the following: Figure 1. An EBIDS detection rule. Each rule is composed of some number of signatures in a particular format. False-positive rate 120 Rate (percent) 100 1.92 3.57 24.55 9.13 75.45 90.87 Miss rate Hit Rate 80 60 40 20 0 EBIDS SpamAssassin Tools Figure 2. EBIDS and SpamAssassin performance numbers. Hits, misses, and false-positive percentages are used as metrics to evaluate an IDS’s effectiveness. The following rule set is used for testing: • Account compromise. Detects when an attacker falsely claims that a user’s account or computer might be hacked, compromised, or otherwise in unauthorized users’ hands. Such attacks can include language that intimates an account compromise, a corporation name to make the entire thing sound official, a threat to suspend a user’s account, or an information request asking the user to follow a link or reply to the e-mail with account information. • Financial opportunity. Detects when a windfall offer is being sent, clearly spelling out a monetary value. Such e-mails typically use congratulatory language to convince recipients that the offer is exciting; money lan104 guage, dollar amounts, or the word “free”; a corporation name; and an information request. • Account change or update. Detects requests for account verification due to a change in a user’s account. Lost account information or account updating are common justifications for requesting such information. These requests typically include language implying account changes, a corporation name, an information request, and a threat of denial of service with a time limit. • Opportunity. Detects all windfall offers that don’t strictly refer to money. This could include vacation offers, credit card offers, and stock tips. For the purposes of this experiment, I focused mainly on credit cards. Such e-mails included language that refered to credit • Authority. People are more likely to comply when someone with perceived authority makes the request. • Reciprocation. People tend to comply with a request when they believe they’ve been given or promised something of value. • Consistency. People will sometimes comply with a request if they’ve made a public commitment or endorsed a cause. • Social validation. People will tend to comply with a request when it appears to be in line with what others are doing. • Scarcity. People will comply with requests when the sought-after object is in short supply, valued by others, or only available for a limited time. TESTING To define the project’s corpus, I divided it into known-bad, a publicly available phishing corpus from Jose Nazario (http://vx.netlux.org/lib/ anj01.html), and known-good, a sub________ division of the publicly available Enron corpus that was absent of phishing e-mails. I then divided these two corpuses into testing and training sets to develop the signatures with some of the corpuses’ literal language, since language evolves and the data being tested doesn’t necessarily use the same language as when I researched EBIDS. The testing and training sets were mutually exclusive. The training set wasn’t used to test the system and vice versa. I used other tools to compare the evaluation of EBIDS results. I installed SpamAssassin with an updated signa- Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page ture set. I also installed Snort and updated it with the most recent rules, but I couldn’t produce any meaningful results. It might have been limitations on my end, so I’m not claiming that Snort can’t be used to detect social engineering, but I couldn’t use it, perhaps due to the system’s nature. Summary of results Several metrics have been established to evaluate the effectiveness of an IDS, as Figure 2 shows They include the hit rate, the percentage of malicious e-mails correctly identified; the false-positive rate, the percentage of innocuous e-mails incorrectly flagged as malicious; and the adjusted false-positive rate, the percentage of false positives plus bad traffic that was correctly flagged for the wrong reasons. As Table 1 shows, with only four detection rules written, EBIDS achieved a relatively high hit rate of 75 percent. That’s somewhat lower than SpamAssassin’s 90 percent hit rate; however SpamAssassin’s signature set is more extensive, requiring 783 Kbytes of storage and comprising thousands of signatures. EBIDS didn’t outperform SpamAssassin as far as detecting malicious social engineering e-mails as they occur. However, the system outperformed SpamAssassin when it came to false positives, with both systems faring relatively well in that regard. False positives were expected to be high with EBIDS because of the nature of its algorithm as it relates to stringmatching plain English. Problems with EBIDS A few common foibles led to higherthan-expected EBIDS miss rate (inverse of the hit rate). One such problem was a subset of the Nazario phishing corpus that included e-mails with subject lines containing the actual misleading natural language and bodies composed entirely of an image or executable/zipped text. I foresaw this problem, but saw more of that type of e-mail than expected. It’s probably possible to cor- A BEMaGS F Table 1. Summary of EBIDS and SpamAssassin results. Metric EBIDS E-mails Known bad e-mails Non-SE e-mails Alerts Correct alerts Completely false alerts Partially false alerts Adjusted false alerts Hit rate False-positive rate Adjusted false-positive rate 536 224 312 175 169 6 71 77 0.75446 0.01923 0.14365 rect the problem by checking the e-mail subject line and testing for blank e-mail bodies that include only executable/zipped text or images. EBIDS also partially false-fired on a specific PayPal security audit fraud, and since that fraud was the most frequently occurring e-mail in the Nazario corpus, the adjusted falsepositive rate was higher than expected. Of course, due to the lowerthan-expected false-positive rate, the adjusted rate just about leveled out to expectation with the high adjusted false positives. Comparison with SpamAssassin SpamAssassin performed much better in its hit rate. It caught 90 percent of the phishing e-mails, looking for technical points in each e-mail, such as the structure of the e-mail code, the legitimacy of the sender and date fields, and other such criteria. The system is finely tuned to detect spam as it comes in, and the e-mails in the corpus aren’t extremely recent, so it fared as well as, if not better than, expected. However, its false-positive rate was higher than expected, nearly doubling the rate of EBIDS. In perspective, the rate was still manageable at approximately 3 percent, which isn’t bad for most practical applications. Overall, SpamAssassin would be the better choice to run in a production environment, even if the current iteration of EBIDS was feasible in production. However, with a more SpamAssassin 549 241 308 230 219 11 0.90871 0.03571 complete signature set, EBIDS might compare more favorably in the future. T he EBIDS project has proven that intrusion detection can effectively detect social engineering in e-mails. Although ripe for refinement, with only four rules developed, the system performed nearly as well as the industry standard. It could eventually let operators put network defense in the hands of a system, rather than users. ■ Allen Stone is a software engineer at Jacob & Sundstrom. Contact him at astone@jsai.com. ___________ Editor: Jack Cole, US Army Research Laboratory’s Network Science Division, ____________ jack.cole@ieee.org; http://msstc.org/cole Renew your IEEE Computer Society membership today! www.ieee.org/renewal 105 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F HOW THINGS WORK SMS: The Short Message Service Jeff Brown, Bill Shipman, and Ron Vetter University of North Carolina Wilmington Although it is a widely used communication mechanism for cell phone users, SMS is far more than just a technology for teenage chat. C ell phones have become an integral part of the modern world, providing human connectivity in a way never before possible. A recent United Nations report (www.cellular-news. com/story/25833.php) estimated that _______________ the total number of mobile phone subscribers in the world now exceeds 2.68 billion. Around 80 percent of the world’s population has mobile phone coverage, with 90 percent coverage forecast by 2010 (www.textually.org/ textually/archives/2006/10/013841. ___________________________ htm). ___ While most cell phones are used for their original intent—making telephone calls wirelessly—these devices are also loaded with other features that are often little used or even ignored. One feature that users have begun to fully exploit in recent years is the short message service or text messaging. This basic service allows the exchange of short text messages between subscribers. According to Wikipedia, the first commercial short text message was sent in 1992 (http://en. wikipedia.org/wiki/Short_message_ ___________________________ service). _____ Although the popularity of text messaging is well established in many countries, in others, such as the US, 106 interest in the thumb-driven phenomenon has only recently skyrocketed. Consider these statistics from the Cellular Telecommunications & Internet Association (CTIA), the international association for the wireless telecommunications industry: • In 1995, roughly 13 percent of the US population had cell phones; by June 2007, it was 80 percent. • In 2000, 14.4 million text messages were sent per month; by June 2007, the number had increased to 28.8 billion per month, which represents a 130 percent increase over June 2006. • In the second quarter of 2007, Verizon Wireless alone says it handled 28.4 billion text messages. Based on these numbers, it is evident that SMS messaging is becoming a widely used communication mechanism for cell phone users. The “Typical SMS Applications” sidebar lists the variety of uses for this technology. THE SMS SPECIFICATION SMS technology evolved out of the Global System for Mobile Communications standard, an internationally accepted cell phone network specification the European Telecommunications Standards Institute created. Presently, the 3rd Generation Partnership Project maintains the SMS standard. SMS messages are handled via a short message service center that the cellular provider maintains for the end devices. The SMSC can send SMS messages to the end device using a maximum payload of 140 octets. This defines the upper bound of an SMS message to be 160 characters using 7-bit encoding. It is possible to specify other schemes such as 8-bit or 16-bit encoding, which decreases the maximum message length to 140 and 70 characters, respectively. Text messages can also be used for sending binary data over the air. Typically, specific applications on the phone handle messages that contain binary data—for example, to download ring tones, switch on and off animation, exchange picture messages, or change the look and feel of the handset’s graphical user interface. The system can segment messages that exceed the maximum length into shorter messages, but then it must use part of the payload for a user-defined header that specifies the segment sequence information. SMSCs operate in either a storeand-forward or a forward-and-forget paradigm. In the store-and-forward paradigm, the system resends the message for some period of time until it is successfully received. In a forwardand-forget paradigm, the system sends the message to the end device without assurance of receipt or an attempt to redeliver in the case of failure. The SMS protocol stack comprises four layers: the application layer, the transfer layer, the relay layer, and the link layer. Table 1 provides an example of a transfer protocol data unit. SMS CONCEPTS AND SYSTEM ARCHITECTURE To fully understand how the various components in the SMS system architecture interact, it is worthwhile to first discuss a few SMS concepts and the role of several independent system entities. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Short codes A short code is an abbreviated number (four, five, or six digits) that is used as an “address” for text messages. Individual carriers can use short codes that are valid only on their network or are interoperable across network carriers (known as a common short code). For example, a carrier might need to send a text message to a subscriber concerning changes in policy or phone configuration, and it probably would send such a message from a short code only that carrier uses. On the other hand, common short codes can be used to send messages to and from users on multiple cellular networks. In the US, one entity, the CTIA (www. usshortcodes.com), adminis__________________ ters common short codes. Content providers can lease a common short code, thereby allowing subscribers to text a keyword to the provider’s short code, and the provider can respond with information specific to the keyword used. For example, texting the word “movie” to short code 90947 might result in a text message being returned to the sender that contains a list of movies showing at a local theater. Google offers a variety of interactive SMS applications through its short code 466453 (http://google.com/sms). _______________ Typical SMS Applications CONSUMER APPLICATIONS • • • • Person-to-person messaging: (chat with friends) Interactive information services: (text to get today’s weather forecast) Entertainment services: (download a ring tone) Location-based services: (restaurant suggestions based on handset location) CORPORATE APPLICATIONS • Notification and alert services: (emergency broadcast messages) • Managing contacts, correspondence, and appointments: (SMS integration with Microsoft Outlook) • Vehicle location: (bus tracking) CELL OPERATOR APPLICATIONS • Subscriber identity module updates—network operators can remotely update data stored on a mobile phone’s SIM card: (remotely update customer service profiles or address book entries) • WAP push—push a URL of the content to be displayed, and a browser on the mobile phone presents the content to the subscriber: (push a mediarich advertisement to a user and automatically display it within the mobile phone’s browser) Because short codes also have premium billing capabilities, wireless subscribers can send text messages to short codes to access and pay for a variety of mobile content and services. For example, if a user sends a text to get a daily horoscope, a charge for the content will appear on the next phone bill. E-mail gateways In the US, most messages sent to applications or services are addressed to a short code instead of a standard phone number. It’s possible to offer a service through a cellular modem, which is, in effect, a cell phone hooked to a computer. However, cellular modems are quite slow and not Table 1. Transfer protocol data unit: 07917283010010F5040BC87238880900F10000993092516195800AE8329BFD4697D9EC37 Octet(s) Description 07 91 72 83 01 00 10 F5 Length of the SMSC information (in this case, 7 octets). Type-of-address of the SMSC (91 means international format of the phone number). Service center number (in decimal semi-octets). The length of the phone number is odd (11), so a trailing F has been added to form proper octets. The phone number of this service center is “+27381000015.” First octet of this SMS-DELIVER message. Address length. Length of the sender number (0B hexidecimal = 11 decimal). Type-of-address of the sender number. Sender number (decimal semi-octets), with a trailing F (“+27838890001”). Protocol identifier (00 = SME-to-SME protocol—implicit). Data coding field (00 = 7 bit, 01 = 8 bit, 10 = 16 bit, 11 = reserved). Time stamp (semi-octets) in order (YY, MM, DD, HH, MM, SS, TIMEZONE in relation to GMT in units of 15 minutes). So, 0x99 0x30 0x92 0x51 0x61 0x95 0x80 means 29 Mar 1999 15:16:59 GMT+2. User data length: length of message. The data coding field indicated 7-bit data, so the length here is the number of septets (10). If the data coding field were set to indicate 8-bit data or Unicode, the length would be the number of octets (9). Message “hellohello,” 8-bit octets representing 7-bit data. 04 0B C8 72 38 88 09 00 F1 00 00 99 30 92 51 61 95 80 0A E8329BFD4697D9EC37 More details can be found at www.dreamfabric.com/sms 107 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F HOW THINGS WORK SMS SMPP SMS SMPP API SMS broker Content server and software applications Aggregator Content provider SMPP SMS Mobile users SMSC Service providers Figure 1. SMS system architecture. Rather than communicating directly with the various SMSCs, content providers go through a message aggregator.The message aggregator uses the SMPP to maintain connections with carrier networks. intended for production environments requiring any significant volume. Most carriers provide e-mail-to-SMS gateways that receive e-mail messages and convert them to SMS messages. For example, an e-mail addressed to 11235551234@mycarrier.com ______________________ could trigger an SMS message for 11235551234 on mycarrier’s network. Some companies offer bulk SMS services and deliver their messages through e-mail-to-SMS gateways. However, not only are these e-mail gateways one-way and unreliable, but carriers also actively monitor them and can block them at any time. Some gateways automatically block messages originating from an IP address when volume reaches a certain threshold. This could cause serious problems in the event of mass text messages being sent for emergency notification purposes. This is a store-and-forward operation. The SMSC usually has a configurable time limit for how long it will store the message, and users can usually specify a shorter time limit if they want. Figure 1 depicts the general system architecture for sending and receiving SMS messages. Each service provider operates one or more SMSCs. Mobile users’ SMS messages are sent through a wireless link via a cell tower to the SMSC. The SMSC access protocols enable interactions between two SMSCs or interactions between external shortmessage entities (SMEs) and an SMSC. SMEs are software applications on network elements (such as a mobile handset) or hardware devices that can send and receive short messages. The SMS Forum has adopted the Short Message Peer-to-Peer (SMPP) protocol to enable interactions between external SMEs and service centers that the different manufacturers maintain and operate. SMS centers When a user sends a text message to another user, the phone actually sends the message to the SMSC, which stores the message and then delivers it when the recipient is on the network. 108 Message aggregators Typically, content providers do not communicate directly with the various SMSCs, but instead go through a broker, or message aggregator, as Figure 1 shows. An aggregator is a business entity that negotiates agreements with network providers to act as a middleman providing access to a cellular network for messaging services to third parties who have no direct relationship with the cellular network. The message aggregator uses the SMPP to maintain connections with carrier networks. Aggregators typically provide access to their servers either through SMPP or using custom APIs written in Java, PHP, Perl, and so on. Most aggregators will also manage the rental of the common short code for clients who do not want to deal directly with the CTIA. Another important function of the aggregator is to assist with provisioning of the short code on the various carriers. Before a carrier allows the use of a short code on its network, it requires a complete description of how the code will be used, including how subscribers will opt in and opt out of subscription services. The provisioning process can take eight weeks or more. Content providers A mobile content provider is an entity that provides value-added content and applications for mobile devices. For example, when a mobile phone user sends an interactive text message to retrieve information, the content provider returns the information (in this case, a text message back to the user) through the aggregator. The aggregator is responsible for transmitting the message to the end user. This second transmission actually masks two transmissions: first a transmission from the aggregator to the cellular provider, and then a transmission from the cellular provider to the mobile handset. BUILDING APPLICATIONS We have established the hardware, software, and network infrastructure necessary to build and deploy advanced SMS applications. We have registered a short code and established a new company, Mobile Education, to develop two-way SMS-based applica- Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page tions. We have formed this company with the expressed intent of developing and marketing intellectual property that we are creating in partnership with the University of North Carolina Wilmington. Mobile Education (www.mymobed. com) currently hosts the following ___ SMS-based applications for UNCW students: • a subscription service for daily campus activities and events; • an interactive system to request and receive real-time shuttle bus information; • an application that integrates with the student Banner system to allow students to request and retrieve a grade for a particular course; • an interactive service to get campus activity and event information, specifically movie dates and times at the campus theater; and • a broadcast text-message-based alert system for campus-originated emergency information. We have learned some valuable lessons while establishing this new business. Vanity code or random code The process of obtaining a short code is different from country to country. In the US, obtaining a short code requires choosing between a “vanity” and a “random” code. Short codes that spell a name, such as Google = 466453 or Hawks = 42957, are known as vanity short codes. Vanity codes cost $1,000 per month. Random short codes, which might or might not spell a word, generally cost $500 per month. Being only five- or six-digit numbers, random short codes are still fairly easy to remember, so it might or might not be worth the extra $500 per month cost to get a vanity code. You can find out who has registered many of the short codes assigned in the US via the Common Short Code Administration Directory (www. ____ usshortcodes.com) or the US Common ____________ Short Code WHOIS Directory (www. ____ usshortcodeswhois.com). _________________ Fees and the provisioning process The fee to lease the common short code simply guarantees that no one else can use that code. A carrier must provision a code before the code can be used on its network. The provisioning process is complicated, expensive, and takes up to eight weeks. During the provisioning process, the carriers examine how their users will interact with the applications associated with the short code. They are particularly interested in applications that involve users subscribing to receive messages on a regular basis. They require double opt-in, which means that the user requests the subscription and then confirms. Users must be able to send a message with the keyword “stop” to indicate that they do not want to receive any more messages. Maintaining a common short code is an expensive, long-term commitment and should be thoroughly researched. The cost of provisioning a short code will generally be between $2,000 and $4,000, including aggregator and carrier setup fees. The monthly fee to maintain service with an aggregator is generally between $1,000 and $1,500. Working with aggregators There are significant differences between aggregators, and you should look closely at more than one before choosing. Several companies that sell aggregator services do not actually maintain SMPP connections with carrier SMSCs, but go through another aggregator instead. The custom APIs that aggregators provide differ widely—some require a constant socket connection, while others use XML over HTTP and do not rely on a constant connection. Some aggregators have relatively inexpensive testing programs and allow you to test their API on a demonstration A BEMaGS F short code. This means that you can send messages that come from the demo code, and if a message sent to that code starts with a special keyword established for you, then the message is routed to your server. However, some aggregators do not allow any testing with their servers until you have signed a contract. Message sending rates There are also vast differences in the rate at which you can send messages. Some plans allow only two messages per second, and some claim rates of up to 100 per second. Some aggregators allow you to send a message to users knowing only their phone number, while some require that you also know their service provider. Carrier rules and restrictions Another important aspect of common short codes is that the mobile content providers who register them must agree to a long list of network operator rules and restrictions. For example, the following is a statement for text message content from a typical short code registration agreement: The following content is not allowed: adult (swimsuit pictures are ok), or any unlawful, harmful, threatening, defamatory, obscene, harassing, or racially, ethically or otherwise objectionable content. Services that facilitate illegal activity, gambling, promote violence, promote discrimination, promote illegal activities, or incorporate any materials that infringe or assist others to infringe on any copyright, trademark, or other intellectual property rights are prohibited. In many cases, network operators even provide lists of inappropriate words that cannot be used within the content that traverses their networks. They can decide to shut down your short code if you breach the agreement you signed to gain access to their network. These restrictions are considerably different from the Internet. 109 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page One more reason to become an IEEE Computer Society member IEEE COMPUTER SOCIETY e-learning campus Advance your career and improve your knowledge with online resources Further your career or just increase your knowledge The e-Learning campus provides easy access to online learning materials to IEEE Computer Society members. These resources are either included in your membership or offered at a special discount price to members. Online Courses Over 1,300 technical courses available online for Computer Society members. IEEE Computer Society Digital Library The Digital Library provides decades of authoritative peer-reviewed research at your fingertips: Have online access to 25 society magazines and transactions, and more than 1,700 selected conference proceedings. Books/Technical Papers Members can access over 500 quality online books and technical papers anytime they want them. IEEE ReadyNotes are guidebooks and tutorials that serve as a quick-start reference for busy computing professionals. They are available as an immediate PDF download. Certifications The CSDP (Certified Software Development Professional) is a professional certification meant for experienced software professionals. Brainbench exams available free for Computer Society members, provide solid measurements of skills commonly requested by employers. Official Brainbench certificates are also available at a discounted price. Visit http://computer.org/elearning for more information Computer A BEMaGS F HOW THINGS WORK Policy issues Maintaining a common short code requires working very closely with an aggregator and being bound by the limitations of its policies and software. It’s an expensive, long-term commitment and should be thoroughly researched. T he short message service has emerged as one of the most popular wireless services. SMS is far more than just a technology for teenage chat. Mobile marketing campaigns are already a very profitable business and growing rapidly. Next-generation SMS applications will incorporate locationbased capabilities that are now being incorporated into mobile handsets. This will enable a new set of innovative services that are targeted and personalized, further refining mobile advertising models and driving revenue growth for carrier operators, aggregators, and mobile content providers. ■ Jeff Brown is a professor of mathematics and statistics at the University of North Carolina Wilmington and cofounder of Mobile Education. Contact him at brownj@uncw.edu. ____________ Bill Shipman is a graduate student in computer science and information systems at the University of North Carolina Wilmington. Contact him at _______ wjs6797@ uncw.edu. _______ Ron Vetter is a professor of computer science at the University of North Carolina Wilmington and cofounder of Mobile Education. Contact him at ______ vetterr@ uncw.edu. _______ Computer welcomes your submissions to this bimonthly column. For additional information, or to suggest topics that you would like to see explained, contact column editor Alf Weaver at ________ weaver@cs. virginia.edu. ________ Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F SOFTWARE TECHNOLOGIES Conquering Complexity Gerard J. Holzmann, NASA/JPL Laboratory for Reliable Software In complex systems, combinations of minor software defects can lead to large system failures. I n his book Normal Accidents: Living with High-Risk Technologies (Princeton University Press, 1984), sociologist Charles Perrow discussed the causes of failure in highly complex systems, concluding that they were virtually inevitable. He argued convincingly that when seemingly unrelated parts of a larger system fail in some unforeseen combination, dependencies can become apparent that couldn’t have been accounted for in the original design. In safety-critical systems, the potential impact of each separate failure is normally studied in detail and remedied by adding backups. Failure combinations, though, are rarely studied exhaustively; there are just too many of them, and most have a low probability of occurrence. A compelling example in Perrow’s book is a description of the events leading up to the partial meltdown of the nuclear reactor at Three Mile Island in 1979. The reactor was carefully designed with multiple backups that should have ruled out what happened. Yet, a small number of relatively minor failures in different parts of the system conspired to defeat those protections. A risk assessment of the probability of the scenario that unfolded would probably have con- cluded that it had a vanishingly small chance of occurring. No software was involved in the Three Mile Island accident, but we can draw important lessons from it, especially in the construction of safety-critical software systems. SOFTWARE DEFECTS We don’t have to look very far to find highly complex software systems that perform critically important functions: The phone system, fly-by-wire airplanes, and manned spacecraft are a few examples. The amount of control software needed to, say, fly a space mission is rapidly approaching a million lines of code. If we go by industry statistics, a really good—albeit expensive—development process can reduce the number of flaws in such code to somewhere in the order of 0.1 residual defects per 1,000 lines. (A residual defect is one that shows up after the code has been fully tested and delivered. The larger total number of defects hiding in the code are often referred to as the latent defects.) Thus, a system with one million lines of code should be expected to experience at least 100 defects while in operation. Not all these defects will show up at the same time of course, and not all of them will be equally damaging. BIG FAILURES OFTEN START SMALL Knowing that software components can fail doesn’t tell us any more about a software system than knowing that a valve can get stuck in a mechanical system. As in any other system, this is only the starting point in the thought process that should lead to a reliable design. We know that adding fault protection and redundancy can mitigate the effects of failures. However, adding backups and fault protection inevitably also increases a system’s size and complexity. Designers might unwittingly add new failure modes by introducing unplanned couplings between otherwise independent system components. Backups are typically designed to handle independent component failures. In software systems, they can help offset individual software defects that could be mission-ending if they strike. But what about the potential impact of combinations of what otherwise would be minor failures? Given the magnitude of the number of possible failure combinations, there simply isn’t enough time to address them all in a systematic software testing process. For example, just 102 residual defects might occur in close to 104 different combinations. This, then, opens up the door to a “Perrow-class” accident in increasingly complex software systems. As before, the probability of any one specific combination of failures will be extremely low, but as experience shows, this is precisely what leads to major accidents. In fact, most software failures in space missions can be reasoned back to unanticipated combinations of otherwise benign events. Minor software defects typically have no chance of causing outright disaster when they occur in isolation, and are therefore not always taken very seriously. Perrow’s insight is that reducing the number of minor flaws can also reduce the chances for catastrophic failures in unpredictable scenarios. 111 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F SOFTWARE TECHNOLOGIES CODING STANDARDS We can reduce the number of minor defects in several ways. One way is to adopt stricter coding rules, like the ones I described in this column in “The Power of Ten: Rules for Developing Safety-Critical Code” (Computer, June 2006, pp. 95-97). Among these recommendations is the required use of strong static source code analyzers such as Grammatech’s CodeSonar, or the analyzers from Coverity or Klocwork, on every build of the software from the start of the development process. Another is to always compile code with all warnings enabled, at the highest level available. Safety-critical code should pass such checks with zero warnings, not even invalid ones. The rationale is that such code should be so clearly correct that it confuses neither humans nor tools. ing more than 600,000 lines of code (K. Grimm, “Software Technology in an Automotive Company—Major Challenges,” Proc. 25th Int’l Conf. Software Engineering, IEEE CS Press, 2003, pp. 498-503). Many robotic spacecraft also use redundant computers, although these typically operate only in standby mode, executing the same software as the main computer. When the controlling computer crashes due to a hardware or software fault, the backup takes over. Adopting stricter coding rules is one way to reduce the number of minor defects. DECOUPLING Another way to reduce Perrow-class failures is to increase the amount of decoupling between software components and thereby separate independent system functionality as much as possible. Much of this is already standard in any good development process and meshes well with code-structuring mechanisms based on modularity and information hiding. Many coding standards for safetycritical software development, such as the automotive industry’s MISRA C guidelines, require programmers to limit the scope of declarations to the minimum possible, avoiding or even eliminating the use of global declarations in favor of static and constant declarations—using, for example, keywords such as const and enum in C. One of the strongest types of decoupling is achieved by executing independent functions on physically distinct processors, providing only limited interfaces between them. High-end cars already have numerous embedded processors that each perform carefully separated functions. Mercedes-Benz S-Class sedans, for example, reportedly contain about 50 embedded controllers, jointly execut112 This strategy, of course, offers limited protection against software defects. In principle, decoupling could be increased by having each computer control a different part of the spacecraft, thereby limiting the opportunities for accidental cross-coupling in the event of a failure. When the hardware for one of the computers fails, the other(s) can still assume its work load, preserving protection against hardware failures. CONTAINMENT The separation of functionality across computers is also a good example of a defect containment strategy. If one computer fails, it affects only the functionality controlled by that computer. Another example of defect containment is the use of memory protection to guarantee that multiple threads of execution in a computer can’t corrupt each other’s address spaces. Memory protection is standard in most consumer systems, but it’s not always used in embedded software applications. For safety- and mission-critical systems, this should probably become a requirement rather than an option. MARGIN Another common defense against seemingly minor defects is to provide more margin than is necessary for system operation, even under stress. The extra margin can remove the opportunity for close calls, whereas temporary overload conditions can provide one of the key stepping stones leading to failure. Margin doesn’t just relate to performance, but also to the use of, for example, memory. It can include a prohibition against placing missioncritical data in adjacent memory locations. Because small programming defects can cause overrun errors, it’s wise to place safety margins around and between all critical memory areas. These “protection zones” are normally filled with an unlikely pattern so that the effects of errant programs can be detected without causing harm. The recent loss of contact with the Mars Global Surveyor spacecraft illustrates the need for these strategies. An update in one parameter via a memory patch missed its target by one word and ended up corrupting an unrelated but critical earth-pointing parameter that happened to be next to the first parameter in memory. Two minor problems now conspired to produce a major problem and ultimately led to the loss of the spacecraft. Had the key parameters been separated in memory, the spacecraft would still be operating today. LAYERING The use of multiple layers of functionality—less gloriously known as workarounds—is yet another way to provide redundancy in safety-critical code. For example, most spacecraft have a nonvolatile memory system to store data; on newer missions, this is typically flash memory. Several independent means are used to store and access critical data on these devices. A spacecraft can often also function entirely without an external file system, in so-called “crippled mode,” by building a shadow file system in main Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page memory. The switch to the backup system can be automated. We can’t predict where software defects might show up in flight, but we can often build in enough slack to maximize the chances that an immediately usable workaround is available no matter where the bugs hide. MODEL-BASED ENGINEERING Perhaps the main lesson that can be drawn from Perrow’s discussion of complex systems is the need to scrutinize even minor software defects to reduce the chances that combinations of minor flaws can lead to larger failures. The strategies mentioned so far deal mostly with defect prevention and defect containment. We’ve left the most commonly used strategy for last: defect detection. Defect detection in software development is usually understood to be a best effort at rigorous testing just before deployment. But defects can be introduced in all phases of software design, not just in the final coding phase. Defect detection therefore shouldn’t be limited to the end of the process, but practiced from the very beginning. Large-scale software development typically starts with a requirementscapture phase, followed by design, coding, and testing. In a rigorous model-based engineering process, each phase is based on the construction of verifiable models that capture the main decisions. Engineering models, then, aren’t just stylized specifications that can be used to produce nice-looking graphs and charts. In a rigorous model-based engineering process, requirements are specified in a form that makes them suitable to use directly in the formal verification of design, for example with strong model-checking tools such as Spin (www.spinroot.com). ______________ Properly formalized software requirements can also be used to generate comprehensive test suites automatically, and they can even be used to embed runtime monitors as watchdogs in the final code, thereby significantly A BEMaGS F strengthening the final test phase (www.runtime-verification.org). _____________________ Finally, test randomization techniques can help cover the less likely execution scenarios so often missed even in thorough software testing efforts. Gerard J. Holzmann is a Fellow at NASA’s Jet Propulsion Laboratory, California Institute of Technology, where he leads the Laboratory for Reliable Software. Contact him at _________ gholzmann@ acm.org. D The research described in this article was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. efect prevention, detection, and containment are all important— none of these techniques should be skipped in favor of any of the others. Even if, for example, developers exhaustively verify all the engineering models used to produce a safety-critical software system, it’s still wise to test the resulting code as thoroughly as possible. There are simply too many steps in safety critical software development to take anything for granted. ■ Editor: Michael G. Hinchey, Loyola College in Maryland; mhinchey@loyola.edu _____________ Giving You the Edge IT Professional magazine gives builders and managers of enterprise systems the “how to” and “what for” articles at your fingertips, so you can delve into and fully understand issues surrounding: • Enterprise architecture and standards • Information systems • Network management • Programming languages • Project management • Training and education • Web systems • Wireless applications • And much, much more … www.computer.org/itpro 113 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ENTERTAINMENT COMPUTING Enhancing the User Experience in Mobile Phones S.R. Subramanya and Byung K. Yi LG Electronics North America R&D munications. Applications-related issues deal with enhancing the users’ experience when interacting with mobile applications. DEVICE-RELATED FACTORS Device-related factors are concerned with features that are built into the device. These features must satisfy several constraints that stem from mobile phones’ small form factor, while also accounting for the devices’ cost and complexity. Improving text input schemes To remain competitive, mobiledevice vendors, developers, and network operators must provide end users with a rich and satisfying experience. W ith more than 2 billion mobile phone subscribers already enrolled, and some 2,000 different mobile phone models in service, mobile communication technology has seen unparalleled growth and penetration. As Figure 1 shows, this technology has swiftly outstripped older rivals to become the most rapidly pervasive and ubiquitous technology to date. However, the average revenue per user for voice communication has decreased steadily, and mobile operators have been increasing their investments in data and entertainment services. In parallel with this trend, the increasing hardware and software capabilities of mobile phones are transforming them from simple voice communication devices to increasingly complex carriers of computing and data services such as messaging, e-mail, and Web access. The entertainment sector has seen comparable growth in the mobile delivery of music, video, TV, and games. Analysts expect both trends to accelerate in the near future, with global revenues from mobile 114 entertainment estimated to reach $37 billion by 2010 (Informa Telecoms and Media; www.informatm.com). _______________ To give some insight into the many functions mobile phones can fulfill, Figure 2 shows the major nonvoice applications used by mobile customers in the US during the first quarter of 2007 (www. ____ mmetrics.com). __________ To drive and sustain this tremendous growth in mobile applications and services, developers must understand mobile users’ needs. In addition to improving core technologies, they must also focus on providing end users with a rich and satisfying mobile-communication experience. The factors contributing to a good user experience include interactions with mobile devices and applications that are natural, intuitive, simple, pleasant, easy to remember, and adaptive to individuals’ idiosyncrasies. Our approach to enhancing the user experience focuses on three dimensions. Device-related issues deal with the hardware features that facilitate ease of use of devices and accessories. Communications-related issues address efforts to enrich person-to-person com- Text messaging is the most popular mobile application, both in terms of subscriber percentage and number of messages sent. To provide a good user experience, manufacturers must develop a means to input text that is easy, quick, error-resistant, and nonstressful. Currently, standard mobile phone keypads require multiple taps when doing text input. The Fastap keypad offers an attractive alternative. Developed by Digit Wireless (digitwireless. com), this technology neatly integrates letters and punctuation keys around the mobile phone’s standard numeric keypad and is available on the LG 6190 phone. This keypad eliminates multiple tapping and is at least twice as fast as a standard keypad. The Sony-Ericsson M600 uses an innovative design that has each key enabling and recognizing distinct pushes—left, right, up, and down—to provide a full QWERTY keyboard using just the keys from a standard mobile phone keypad. This facilitates easy messaging and e-mail. The M600 also features touch screen and handwriting recognition. Special hardware for gaming The LG SV360 has a 1-million polygon-per-second graphics accelerator chip that can process data much faster than standard mobile phones. The enhanced graphics let users play games with more detailed and realistic 3D graphics. The relatively larger 2.2-inch LCD screen lets users play 3D games with improved display quality. The Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page 100 BEMaGS F Television Electricity Radio Penetration of potential world market (percent) A Telephone VCR 80 Automobile 60 PC 40 Internet Cellular 20 0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 Years Figure 1. Mobile technology’s meteoric rise. Mobile phones have proliferated faster than any previous technology. Source: Joseph Jacobsen, Organizational and Individual Innovation Diffusion. Special hardware for music, video, and TV Downloaded mobile game 6.75 Purchased wallpaper or screensaver 6.84 Used work e-mail As another example of mobile phones’ increased versatility, some models now have music players that use simple thumb-wheel scroll buttons to control the player’s volume, eliminating standard models’ awkward navigation across several menus to reach the volume control. Users also can manipulate the thumb-wheel to navigate across menu items of other applications, which is easier than button presses. For example, it could also be used to flip channels in the mobile TV and control video playback. 10.19 Used mobile instant messenger 13.80 Used personal e-mail 17.34 Purchased ringtone 20.00 Browsed news and information 20.48 Used photo messaging 30.67 Sent text message 81.18 0 10 20 30 40 50 Millions of users 60 70 80 90 Figure 2. Mobile applications spectrum.This chart shows nine of the many functions US mobile phones provided during the first quarter of 2007. phone also sports acceleration sensors that let users interact more closely and provides them with better controls, a more realistic feeling, and greater game enjoyment. Pushing this trend still further, Qualcomm is licensing the unified shader architecture incorporated in gaming-console graphics processor units for rendering high-performance graphics in their mobile-station modem chipsets. This fusion will enable even higher-end graphics for gaming applications on mobile phones. Touch-sensitive screens Touch-sensitive screens have been deployed in a few recent phone models, including Apple’s iPhone (3.5” screen with 320 480 pixels resolution), LG’s KE850 (3” screen with 240 300 pixels resolution), and Samsung’s F700 (2.8” screen with 240 440 pixels resolution). The touch screen facilitates provision of a natural interface for many of the phone’s applications. For example, for dialing and receiving calls, it provides the traditional mobile phone keypad; for the calculator application, it displays the 115 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F ENTERTAINMENT COMPUTING typical calculator keypad; and for the music-player application, it displays controls such as the Play/Pause, FastForward, and Rewind buttons. It thus provides the flexibility of a natural interface suited to each application, instead of forcing use of the phone’s keypad to control several different applications. COMMUNICATIONS-RELATED FACTORS Enhancing the mobile-user experience for person-to-person communications broadly involves providing the communicating parties with a feeling of being as close to face-to-face communication as possible. Providing a virtual vicinity Providing video clips and background images of the conversing participants’ surroundings could provide a feeling of sharing the same vicinity. One application developed in this area, SWIS (“see what I see”), lets what one person sees be shared with another as the two converse. This requires voice communication with a concurrent video session that allows content sharing. Techniques must be developed for providing these capabilities costeffectively by judiciously using combinations of available technologies and bandwidths. One simple example might be the appropriate use of prestored images and video clips, drawn from earlier communications. Thus, instead of transmitting this data again, which would consume precious bandwidth, the system can seamlessly pull it from the two parties’ respective mobile phones. Providing feelings of touch Currently, mobile phones provide only the auditory and visual dimensions in communications. A natural next step would be to provide feelings of touch. For example, the development of technologies and techniques for providing a virtual hug, handshake, or kiss— depending on the relationship’s intimacy—would significantly enhance the users’ communications experience. 116 Providing gestures Gestures are unique to individuals. Adding gestures in communications would provide the communicating parties with a sense of closeness. For this to be feasible, researchers must improve the technology for determining the gestures’ parameters, together with schemes for gesture recognition and classification. The gestures could be constructed by receiving only elementary information, such as gesture types and associated parameters, thus using minimal bandwidth. Personalization of many interface facets would help users perceive and react quickly to events on the mobile phone. contexts, varying moods, limited use of hands, and expectations of quick and easy interactions. Avoiding feature overload Mobile applications developers should learn from the mistakes of PC applications developers and pay careful attention to mobile feature sets. In most PC applications, typically 90 percent of users never take advantage of about 90 to 95 percent of their applications’ features. With mobile device resources being limited, the feature set must be minimized and should be derived through extensive use-case studies and diligent analysis. Subsequently, the distilled features should be implemented efficiently. Researchers must develop both quantitative and qualitative metrics of feature-effectiveness and incorporate them into feature design and implementation. Hands-free usage facilities Differentiated communications Communications between the calling and called parties could be differentiated based on their relationships, call times, and contexts. For example, supposing an intimate relationship between two parties, such as a husband and wife, and an appropriate time, such as outside scheduled meetings or business hours, the communications could include visual effects such as avatars, live video, or short video clips of the two parties. APPLICATIONS-RELATED FACTORS In the person-(machine)-to-machine(person) part of mobile communications, the applications’ user interface provides the most important layer directly contributing to the user experience by hiding the complexities of the underlying hardware and software. The UI should adopt a user-centric design that considers several parameters primarily from a user’s perspective. For example, in addition to the mobile devices’ constraints, design should take into consideration mobile-user constraints such as limited attention span while moving, changing locations and The traditional modes of WIMP (windows, icons, mouse, and pointing device) interactions—developed for PC applications—would not work well on mobile devices. Mobile users’ limited attention span, coupled with mobile devices’ limited keypads, call for novel interaction modes. For example, interactions should be carefully designed using judicious combinations of audio and visual cues, along with voice-activated commands. Although several mobile phone models provide voice input, these are restricted to dialing and only a few applications. The provision of voiceactivated commands, which are less error-prone and robust in noisy environments, facilitates increased use of hands-free operations. Personalization Mobile devices are inherently personal. Users would prefer to have unique mobile phone content, such as ring tones, wallpaper, and themes. Many times, users might expect that their devices would provide the facilities for creating such content themselves. Ring-back tones have recently become popular. These are tunes and Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page music set by users, which callers hear before the call is picked up. An extension of this could be that a user could set different ring-back tones for different sets of callers. Personalization of many interface facets would help users perceive and react quickly to events on the mobile phone. Personalization— such as configurable menus based on usage frequencies—also caters to user idiosyncrasies. Byung K. Yi is a senior executive VP at LG Electronics North America R&D. Contact him at _____________ byungkyi@lge.com. A BEMaGS F Editor: Michael van lent, vanlent@ict.usc.edu _____________ Combining visual and auditory stimuli Using combinations of text, audio, images, graphics, and video is an effective means of conveying information. Interfaces should use judicious combinations of visual and auditory stimuli and feedback. As a specific example, associating distinguishing sounds for the different menu selection classes would provide users with auditory feedback that, together with the visual stimulus of icons or text, will likely speed up selections and reduce errors. This is especially beneficial in the mobile context when users’ attention spans are shorter and they cannot afford to stare at the screen for as long as they could while stationary. Any media imbalance or overload should be avoided because it would adversely affect ease of comprehension and use. ______________________ I t is important to consider the user experience while designing mobile devices, applications, and user interfaces. The user experience should take into consideration the constraints of mobile devices as well of those of mobile users. It is crucial to develop models for the description and metrics for the evaluation of the effectiveness of the mobile user experience. This opens up many exciting opportunities for researchers and developers in these areas. ■ S.R. Subramanya is a senior research scientist at LG Electronics North America R&D. Contact him at subra@ lge.com. 117 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F INVISIBLE COMPUTING Taking Online Maps Down to Street Level Luc Vincent, Google Street View enables simple navigation between street-level images without losing the map context. To address this limitation, Google launched the Street View feature of Google Maps in May 2007. The underlying idea is very simple: Provide an interface that can display street-level images in a natural way that enables convenient navigation between images without losing the map context. Figure 1 shows a screen shot of the interface. To use Street View, you simply point your browser to _____ http:// maps.google.com, visit one of the 15 _____________ cities that currently has coverage, such as San Francisco or Portland, and move around by either clicking on the map or navigating from image to image. EARLY RESEARCH EFFORTS S ince MapQuest.com’s __________ creation in 1996, online mapping systems have rapidly gained worldwide popularity. With Google Maps leading the charge, new sophisticated AJAX (Asynchronous JavaScript and XML) mapping applications began appearing in 2004. These geographic Web 2.0 applications, which continually add new features and improvements, have made online maps almost as essential to our daily lives as the search engine. Figure 1. Google Maps Street View interface. 118 A fairly recent offering of the best online mapping systems is aerial imagery. Such imagery takes the so-called GeoWeb to the next level, providing tremendous value to activities such as real estate sales, insurance, environmental management, municipal government, emergency services, and law enforcement, as well as engaging serendipitous Web users. Yet the imagery often is not detailed enough, and viewing buildings and streets from above can be disconcerting. This idea behind Street View is not new. The earliest related project dates back to the late 1970s, when MIT’s Andrew Lippmann created a system known as Movie Maps that let users take virtual tours of the city of Aspen, Colorado (A. Lippman, “Movie Maps: An Application of the Optical Videodisc to Computer Graphics,” ACM SIGGRAPH Computer Graphics, vol. 14, no. 3, 1980, pp. 32-42). The Movie Maps data-capture system used a gyroscopic stabilizer and four cameras mounted atop a car, with an encoder triggering the cameras every 10 feet. The system digitized the captured panoramic imagery, organized it into several scenes, each covering approximately a city block, and stored it on a laser disk. A user could then virtually navigate Aspen streets using the interface shown in Figure 2. Several related systems subsequently emerged, with the pace of development accelerating in the late 1990s and early 2000s. One noteworthy offering was BlockView in Amazon’s A9.com ______ search engine. Introduced in 2005, BlockView was the first Web-based system to expose large amounts of street imagery to users. However, it was discontinued less than two years later. More recently, the Street View system itself went through several phases before going online. Google cofounder Larry Page bootstrapped the project by Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Figure 3. Example pushbroom perspective image. Figure 2. Movie Maps interface. personally collecting a sample of urban video footage from his car using a camcorder. This indirectly led to the idea of creating long, seamless images with somewhat unusual perspectives known as “pushbroom panoramas,” an example of which is shown in Figure 3. In 2006, then Stanford University graduate student Augusto Román developed techniques to create more visually pleasing “multiperspective panoramas” (A. Roman, “Multiperspective Imaging for Automated Urban Imaging,” doctoral dissertation, Dept. Electrical Eng., Stanford University, 2006, ______________ http:// graphics.stanford.edu/ papers/aroman_ thesis). _________________________ However, such imagery was difficult to generate and perhaps even harder to use in an intuitive user interface, so 360-degree panoramas took center stage and ultimately led to the current implementation. The Street View team also tied the imagery to the map in a way that makes navigation easy, whether from the map or from the imagery itself. In particular, the Street View avatar, often referred to as “pegman,” rotates and moves along with the map even when navigating from the panoramic image itself. However, Street View is not merely an exercise in advanced interface design. To collect and process the vast amounts of imagery required for the application, a fleet of custom vehicles such as the one shown in Figure 4 is currently deployed and busily capturing data that will enable coverage well beyond the current 15 US cities. According to The World Factbook: 2007, there are more than 19.4 million kilometers of paved roads around the globe, so the Street View data-collection team expects to be busy collecting data for quite a while. OTHER CHALLENGES Aside from interface design and data collection, Street View poses challenges related to fleet management, pose optimization, and computer vision. Fleet management When more than just a few cars are involved, there is a strong financial incentive to manage them as efficiently as possible. Fleet management is an active area of operations research and of great interest to the transportation industry (T.G. Crainic and G. Laporte, eds., Fleet Management and Logistics, Springer, 1998). Managers and dispatchers can use operations research techniques to optimize the deployment of a fleet of data-acquisition vehicles. Using local planning techniques can also ensure that drivers traverse each road segment as few times as possible, which is difficult given road graph constraints (L. Kazemi et al., “Optimal Traversal Planning in Road Networks with Navigational Constraints,” Proc. 15th ACM Int’l Symp. Advances in MERGING IMAGERY WITH MAPS Street View literally puts users in the driver’s seat. One of the key technical challenges Street View addresses is the merging of street-level imagery with maps in a natural way. In addition, like most Web 2.0 applications, it works across numerous operating systems and browser combinations. Yet, Street View is more than a pure AJAX application in that it adds Adobe Flash technology to the mix: The “picture bubble” is actually a Flash application that delivers a rich graphic experience without requiring use of a custom browser plug-in. This richness is best experienced by rotating the view right or left 360 degrees or zooming in on details such as parking signs. Figure 4. Street View custom data-collection vehicle. 119 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F INVISIBLE COMPUTING Geographic Information Systems, ACM Press, 2007). Pose optimization When capturing imagery, it is essential to know as accurately as possible each image’s location and the camera’s orientation. However, the Global Positioning System alone is typically insufficient to provide this data, especially in cities with many “urban canyons,” streets flanked with dense tall buildings. Our solution is to equip the vehicles with additional sensors such as rate gyros, accelerometers, and wheel encoders to capture vehicle speed and acceleration at a high sampling rate. Software such as James Diebel’s Trajectory Smoother (http://ai.stanford. edu/~diebel/smoother.html) or the ___________________ open source Google Pose Optimizer (http://code.google.com/p/gpo) can then be used offline to compute an improved pose estimate from these measurements. 120 Computer vision Creating 360-degree panoramas by stitching together a series of digital pictures is not a new problem, and many excellent panorama-stitching packages are available such as Autostitch (M. Brown, “Autostitch: A New Dimension in Automatic Image Stitching,” www.cs.ubc.ca/~mbrown/autostitch/ autostitch.html). ___________ However, it is one thing to manually create a few panoramas and quite another to automatically generate millions of attractive panoramas across a wide variety of lighting conditions. Solutions involve a pipeline of complex algorithms for camera calibration, vignetting correction, color correction and balancing, image alignment, stitching, blending, and so on. A project like Street View requires complex engineering but surprisingly little groundbreaking research. What makes it uniquely chal- lenging is its intrinsic scale. This probably explains why few of the dozens of similar projects have reached Street View’s still rather modest size. However, as more users and organizations appreciate the value of street-level imagery, similar services are sure to emerge and accelerate innovation in this space. Such imagery will become increasingly integrated with the overall online map experience and might extend beyond streets to include interiors, real estate, national parks, and more. In fact, not even the sky is the limit: The Street View panoramic viewer is featured in the recently updated Google Moon service (http://moon.google. com) to showcase pictures taken dur___ ing various Apollo missions. ■ Luc Vincent leads several geographicrelated projects at Google. Contact him at ____________ luc@google.com. Editor: Bill N. Schilit, Google; bill.schilit@computer.org, ________________ Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F _______________ ____________________________ Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F THE PROFESSION Continued from page 124 external devices that can easily be attached to the computer, but that the computer can do without. We should make the operating system as efficient as possible so that, contrary to the rather vulgar tendency these days, it can run on slower rather than faster CPUs. Good programs should run at an acceptable speed on a 50-MHz CPU. A slower CPU means less power consumption and a longer battery life that, if the computer is used for work, we need a lot more than semitransparent menus or fading shadows below and on the right of windows. Frankly, assuming that computers these days should need a 1-GHz machine to make a menu appear in a reasonable time is ridiculous. DO WE REALLY NEED COLOR? We are not awestruck teenagers, so we don’t need shaded and embossed windows. Simple, possibly black-andwhite windows would be more than adequate. I am not an expert in these areas, but if a black-and-white screen can save some power, add a couple of hours to battery life, and take away a few pounds, then I welcome it. For people who absolutely need color, it shouldn’t be hard to add this option to the operating system. The designer could also add a switch to turn off the computer. A real one, I mean, with two positions separated by a click. I am tired of unplugging the computer or the battery every time the operating system crashes disastrously. On the other hand, we must avoid building the computer too small: If the intended user does a lot of writing, a small screen becomes a pain, a small keyboard even more so. I usually carry many papers, so I have no incentive to make the computer smaller than an A4 or letter-size sheet. All the simplification a simpler operating system affords should be directed toward making the computer very thin and light. DIE, INSTALLATION, DIE! Installing a program should only require copying a bunch of files in a directory; uninstalling it should only require removing those files. This is 122 how a user installs emacs (http:// ____ directory.fsf.org/project/emacs), which provides a good model for how things should be done. There should be no dependencies between programs. Few things annoy more than installing program A only to discover program B must be installed first, which in turn requires program C, and so on until the user runs out of patience, letters of the alphabet, or both. _____________________ Many computer programs do have shortcuts, but they are too clumsy and limited to be of use. The programs should follow the operating system’s philosophy: Be small and run fast. Using half a Gbyte for a word processor and a few other things is sheer madness. Overloading the program with features few people will ever use is even worse. Interfaces based on menus and buttons might be fine for the casual user, but for the professional, they are too slow. They should be offered as an option, but for the serious user there is nothing faster than keyboard commands that don’t require hand movements, grasping the mouse, and performing complex motions. I challenge anyone to use MS Word to write a mathematical paper faster than I can with LaTeX. Many computer programs do have shortcuts, but they are too clumsy and limited to be of use. MS Word, for example, makes it impossible to typeset an equation or place a figure using just the keyboard. Presentation software presents a delicate issue, with undertones that border on the religious. Edward R. Tufte, in his The Cognitive Style of PowerPoint: Pitching Out Corrupts Within (Graphics Press, 2006), argued cogently that presentation software has degraded the quality of presentations. So the simpler the presentation software the better. Of all the presentations with animations I have seen— including my own youthful efforts— every single one would have been the better for removing the animation and all colors but two. Compatibility should be enforced only when it doesn’t obstruct speed and simplicity. For a word processor, for example, the MS Word .doc format is too cumbersome: Using it condemns the word processor to inefficiency. Using the .rtf or TeX formats will provide all the necessary features with far fewer complications. Unless there is some good reason for it—such as the Internet—compatibility is just a hassle. Let us not forget that any good innovation came about because somebody did things differently from how they were done until then, compatibility be damned. IT’S THE VACCINE THAT KILLS YOU The operating system should prevent programs from loading anything from the Web into the computer unless the user explicitly tries to download such software. Most programs I use these days have the unfortunate mandate to keep downloading upgrades, patches, and fixes without telling me anything, to the point that I now fear the programs I have been forced to install more than I fear any virus. Putting something in my computer without my explicit request is tantamount to trespassing and squatting: In the unfortunate absence of legal measures against software manufacturers, the operating system should block any unrequested download. If this means living without cookies, so be it: I am fairly sure that we will survive. If this means we must live without the razzmatazz of Web sites that look like an LSD flashback, even better. We will definitely be better off without them. H aving described what I would like to see in a computer, I admit that many of its characteristics, such as the search for minimal memory occupancy and maximal speed, might appear to make more sense for portable than desktop computers. Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page Consider, then, the following two points. First, these days more and more people use a portable as their main computer. So, why not design an operating system that, instead of being basically a desktop operating system with which laptops must live, provides a laptop operating system with which desktop computers must live? Second, by forcing programmers to design code focused on maximizing memory occupancy and processing speed, we will get better quality code. Programmers, quite unsurprisingly, seem to produce just that when guided by something other than an impossible delivery schedule and continuous requests for features. This, at least, is the lesson we might be inclined to draw by comparing the quality of, say, embedded programs with the things that usually run on our PCs or, even worse, on the Internet. A simpler operating system architecture, proper algorithm design, and a more careful implementation—one that gives priority to processing speed and memory A BEMaGS F occupancy—would likely create a better operating system. ■ Simone Santini is an associate professor at the Universidad Autónoma de Madrid, Spain. Contact him at ___________ simone.santini@ uam.es. _____ Editor: Neville Holmes, School of Computing, University of Tasmania; neville.holmes@utas.edu.au. Links to __________________ further material are at www.comp.utas. edu.au/users/nholmes/prfsn. __________________ _________________________________ 123 December 2007 Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F THE PROFESSION Making Computers Do More with Less Simone Santini, Universidad Autónoma de Madrid The ideal computer will keep things simple, small, and efficient. M y work these days consists mostly of doing mathematics. I don’t use the computer often, except for writing articles, answering e-mail, and reading the online version of some newspapers. I still do some programming, but recently my students have been doing most of it for me. Still, people know I work with computers and, for most of my friends and relatives, the boundary between a university professor and a system administrator blurs beyond perception. Thus, they often ask me to help solve some complicated computer issue and, being basically a good guy, I try to help them. This is how I came to spend the best part of last weekend helping a friend untangle a rather complicated mess involving, among other villains, a file exchange program, a firewall, a printer driver, and—in the role of the master villain it plays so well—the operating system manufactured by a well-known company headquartered in the northwestern corner of the US. The problem, which involved multiple, overlapping, and frustratingly idiosyncratic interactions among programs, drove me up the wall. Several times I proclaimed, quite loudly and 124 in terms not suitable for American national TV, what these people should do with their operating system. I remember at least once making an oath to abstain forever from any further interaction with computers. Alas, abstain I can’t, since these days I find it impossible to be a professional in any field unless I’m willing to interact with the frustrating machines I incautiously chose to associate myself with professionally. COMPUTING FOR WORKERS Modern programs are disastrously unreliable artifacts, but they needn’t be that way. In no other complex device would we accept a degree of unreliability comparable to that found in software. If we inaugurated a new building only to find that disconnecting the elevator causes all the fifthfloor faucets to turn on, or that changing a light bulb in the basement results in a minor collapse in the east wing, we would probably hang the architect by the toes. That we accept this and worse when it comes to software doesn’t speak well of our profession’s technical culture: We seem to be the first professionals for which quality is a secondary concern. It might not, however, be too late to do something about it. If someone out there itches to design a new operating system, computer system, or application program, I submit my idea for how I would like to see it done. The computer and programs I envision are meant to be for people who use a computer as a work instrument, not as a toy or pastime. When I have some spare time, I walk, attend the theater or cinema, go to a party, or have a drink with friends. If you see me with my hands on a computer keyboard, it means I am working, and when I work I don’t want to be entertained by colorful programs: I want a fast and efficient system to do whatever I need done, in the shortest time possible so I can get on with my life. Alas, these days no major operating system appears designed completely around the guidelines I propose. MAKE IT SIMPLE, MAKE IT SMALL My ideal operating system’s main design guideline mandates keeping things simple, small, and efficient. Eliminate all unnecessary features and then eliminate half of what is left. Einstein said that we should do things as simply as possible, yet no simpler. Fortunately, with computer programs, things can be kept pretty simple and still be functional. First, design a simple, small, fast kernel to deal with secondary memory backup, persistent core memory allocation, and process management, à la Unix, and with some basic network functions. The kernel should be small and easy to install, the rest of the operating system equally so. An operating system should never need to occupy more than 50 Mbytes of secondary storage, but we can grant it twice as much: 100 Mbytes. An operating system this size, with similarly small programs, can run a whole computer on a 500-Mbyte memory stick, with abundant memory to spare, so that somebody can finally build a laptop that weighs less than a pound and has a battery life of 20 hours. Magnetic disks and CDs should be Continued on page 122 Computer Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page ________________________ A BEMaGS F ____________ ____________________________ ____________ ________________ ______________ Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F Mild-Mannered ALM, Super Quality Developing quality software requires a heroic effort, from tracking thousands of the tiniest details, to keeping team communication flowing smoothly. TestTrack Studio 2008 powers the application lifecycle, automating processes and keeping track of issues, change requests, test cases, and test results. With TestTrack Studio 2008, you have the tools and the time to prioritize, communicate, and track the status of your projects more effectively, without breaking a sweat. Let TestTrack Studio 2008 do the heavy lifting—Be a superhero! Download your fully functional evaluation software now from www.seapine.com/ieee08 ©2007 Seapine Software, Inc. Seapine, TestTrack Studio and the Seapine logo are trademarks of Seapine Software, Inc. All Rights Reserved. All other trademarks are the property of their respective owners. Computer Previous Page | Contents | Zoom in | Zoom out | Front Cover | Search Issue | Next Page A BEMaGS F