Telework Collaboration in High-Speed Environments
Executive Summary
Our Efforts have been aimed at three areas:
1. Identify improvements that have been observed and the metrics used to quantify the
improvement due to telework environments and the key factors that affected the results.
Identify work to date that demonstrates the value of real-time, visual environments as key
enablers for collaboration.
2. Assess the state of the access technologies enabling high speed networking, architectures that
are possible with these technologies, key factors that are affecting the deployments and gather
enough information to asses the technical and business environment for the evolution of high
speed access technology. This work then can be leveraged to build a business case model that
can be used with Telework Customers and Partners to address the economics of providing true
teleworking with visual collaboration. During the second quarter of the contract, this report has
been updated to include cable technology for high speed access.
3. Identify tools and systems that have been developed for real-time effective collaboration.
During the second quarter of the contract, we have looked at collaboration systems and efforts
were carried out to create a video conferencing environment using Conference XP.
Collaboration Systems and Recommendation
There are a variety of enterprise-strength community tools that foster increased productivity in a
networked environment. Virtual whiteboarding, among the fastest-growing tools in this space,
encompasses a wide variety of applications, which enable users to write and draw at the same time,
with changes instantaneously visible to all. A key to high performance and acceptability of Telework
is video in conjunction with these distributed software tools to create true virtual meeting rooms and
Telework environments that go across physical boundaries. Furthermore, these boundaries may
extended across the country and international as well.
A major component of collaboration are tools that support the need to meet and share information.
These tools should foster two way communication. A key difference between asynchronous
communications such as email, file exchange, etc. and synchronous communications is that, in a realtime, synchronous communication environment people can build on each other’s ideas and true
collaboration occurs in a way that the capability and results of the team in a team environment is
greater than the individual capabilities one at a time. Collaboration is a result of interaction, and it is
real-time teaming that creates the context for interaction. True real-time collaboration fosters idea
creation. Ideas rarely arise through the inspiration of a genius; more commonly, they arise out of
conversation and interaction—both of which are enhanced in a visual realtime environment
Virtual meeting software enables collaboration and is starting to gain widespread acceptance. Virtual
meeting rooms can include screen sharing, virtual product demonstration, and tools where users can
see, write and draw in common spaces in real time for all to see. Application viewing allows users to
see other’s desktops and the applications they run on it. The advanced collaborative features coupled
with video allow for advanced community interaction to take place, thus simulating real-life physical
meetings. Thus, needed information is shared and conversations can happen quickly and more
efficiently. Leaders in this space include Groove, eRoom, and Webex. However, we observe that
these tools can be enhanced by integration with real-time video capabilities.
Telework and Collaboration with Broadband Access
Page 1 of 36
Some advanced capabilities have been implemented in a corporate intranet environment leveraging
high performance transport since the application with video needs higher bandwidth than is typically
obtainable via dial-up connections. However, in a larger DSL and Cable are needed at the minimum
and two-way symmetric bandwidth for higher quality real-time video. Thus, the efforts of the
Telework Consortium to enable higher bandwidth, symmetrical video enabled communications is
very important to make true teleworking work.
In an effort to build the next generation collaboration environment, we have looked at the market re
what is available and how it can be improved and meet the need, if high bandwidth were available.
Based on the information gathered, we are recommending the creation of a high performance
collaborative work environment which consists of an integrated video system and collaboration tools,
leveraging products available with customization tools. Current Systems do not integrated video
conferencing. The Teleworks consortium is on the way with the Genesis Collaborative application
framework as exemplified in the Internet Video Conferencing (IPVC) from Grant Holcomb at Tulane
University and software from Groove. We are also looking at the Conference XP video conferencing
software, which uses more standard protocols (e.g., RTP). It is currently designed to use servers on
Internet2 and is an evolving research tool from Microsoft, in collaboration with Internet2 users. We
recommend enhancing the current Telework demonstration capability with collaboration systems and
creating a demonstration virtual meeting/workspace environment and demonstrate impact for a work
environment scenario, not just the video conferencing by itself. An integration of the video
capabilities with collaboration software could demonstrate that virtual collaboration can be as good as
collaboration in the physical proximity and even better (since more advanced video, database,
sharing, and other tools will be part of the new environment.) Security mechanisms for this
collaboration environment for an “on demand group”, i.e., groups with changing membership, should
also be integrated and demonstrated.
CORPORATE Deployment Examples
IBM has implemented Lotus’ Sametime instant messaging product on a companywide basis. It is used
several times each business day by more than half of all employees, in effect replacing telephone
communication. Cisco uses instant messaging in similar fashion, as well as Video on IP for Video
communications. IBM also uses an internal “e-meetings” system as an online meeting space, and almost
all major public C-level presentations use e-meetings. The e-meetings software contains application
sharing (so that participants can see PowerPoint or Freelance demonstrations, for example), audio-visual
capabilities, and video-enablement. Skill Tap is one of the newest online collaboration tools at IBM.
Skill Tap enterprise expertise software takes the form of Lotus’ Sametime conversations enabled by
bots. When an IBM Skill Tap user has a business question regarding any issue, he can just send a
standard Sametime message to the Skill Tap bot. The bot then forwards the query to a pool of
experts, and the experts then are expected to return with a result. This process lasts generally
between 15 and 30 minutes. Skill Tap has launched just recently at IBM and it is only now
beginning to be publicized.
SECTION I - Examples of Collaboration Systems
eRoom’s www.eroom.com collaboration package, Digital Workplace, relies on the concept of virtual
teams working in rooms focused on distinct issues. Room templates can be chosen from several
provided or created from scratch. The user can then select members for the group and automatically
have e-mail invitations sent out to confirm their participation. Once the initial process of creating a
room is finished, users have access to a Windows desktop-like interface with customizable buttons to
open specific parts of the room. Information to be shared can be made available to the entire group.
Telework and Collaboration with Broadband Access
Page 2 of 36
Business requirements and goals that need to be met are displayed. Users can also alert others within
an eRoom using e-mail announcements. However, video is not integrated and if it were, then the
interactions would be much facilitated.
eRoom supports integration with a wide variety of corporate applications. Microsoft Office is
supported, so users can drag and drop Office documents and group edit during meetings. Microsoft
Outlook integration allows users to synchronize their calendars and import events to and from eRoom.
eRoom also offers integration with many popular enterprise database programs. Through the use of the
eRoom API Toolkit, eRoom can also be integrated with corporate portals and other commonly used
third-party software.
eRoom’s document management system is easily accessed and users can drag and drop any documents
from Windows folders into the document folder. eRoom provides a complete hierarchy for documents,
which users can modify by dragging and dropping files appropriately. eRoom also provides easily
accessible options for setting file permissions.
There is a Project Calendar, which is shared by all members of an eRoom for meeting scheduling. Any
member of an eRoom can schedule meetings beforehand using the Project Calendar or start at any
time. This calendar can also be integrated with personal calendars. Within the meeting room, eRoom
supports group editing of supported document formats, virtual whiteboards, and desktop sharing.
eRoom also provides online awareness so users can communicate with and see who is online at any
given time and also have real-time conversations, but integration with real-time video support is
lacking. Private chat is also supported in addition to normal group chat features. The Project Inbox
lets users see all e-mail messages related to that eRoom, e.g., meeting invitations and document review
requests.
eRoom’s workflow seems quite well planned. Users can click on a document for review and designate
the people to which it will be forwarded as well as the order in which to notify others, who would be
receiving e-mail alerts asking them to review the document along with a link to the content itself.
After one reviewer has finished editing the document, the workflow system automatically forwards a
link to the edited document to the next reviewer in line. Versioning, and tracking of all changes is
supported.
Groove www.groove.net has tools available for working on various types of content; synchronous
editing of Microsoft Office documents, with many add-ons to expand functionality. Does not appear to
have advanced workflow capabilities; Few options for integration with third-party enterprise software.
However, it is relatively simple with a good user-friendly interface, beyond which lies a powerful
tool with many innovative options and a great deal of customization. Workspace focuses on providing
all necessary business tools within one environment, freeing a user from switching from application
to application to complete his work and uses the concept of a “shared space.” Within this area, a
small group can have access to the appropriate data and the tools needed to work with this content.
Workspace’s interface is fairly intuitive and combines the functionality of many common enterprise
applications. All documents and content are stored locally on users’ desktops, easily accessible to any
member within the space. Through the integrated discussion forum, users can leave messages or
suggest ideas that can then be elaborated or critiqued by fellow employees. Users can choose to fill a
discussion thread in various ways. Integrated video and video conferencing would add much value.
All documents and content within the shared space are encrypted, removing any security concerns.
Nearly every feature in Groove allows users to set permissions dictating the access and modification
of content.
Using the integrated awareness feature, users can see who is active in the shared space and exactly
what content they are viewing or modifying. Groove’s innovative real-time editing tool lets users
Telework and Collaboration with Broadband Access
Page 3 of 36
modify Word documents or view Powerpoint presentations simultaneously, allowing all parties to see
the modifications as they are implemented. Furthermore, employees can communicate through text or
voice in the included chat window. Additional integration with real-time video conferencing can add
value. The content review process is also simple, since it allows employees to select reviewers, set
deadlines, and record any changes made.
Groove Meetings is a scheduling tool that incorporates employee calendars and lets users set up discussion times for anyone within a shared space. Using this last tool, employees can also designate
meeting roles, create agenda items, and synchronize with Microsoft Outlook. The Project Manager
tool lets users create groups to work on specific issues. Using this tool, project organizers can
prioritize tasks, set deadlines, and preserve best practices by saving descriptions of successful projects.
Groove also features several additional third-party tools and features a development environment and
ability to integrate other enterprise data.
Groove Notepad and Outliner are note-taking tools during presentations and meetings. For designoriented work, SketchPad allows users to hand-draw items and modify JPG and BMP images (Real
time Video would make this better). Groove Pinboard serves as an informal reminder and alarm system
and also lets users create polls to gauge opinion within the shared space. Also available for this
purpose is Groove Forms, which allows users to create custom applications for entering and
manipulating data.
iManage’s WorkSite suite by iManage www.imanage.com is a nice enterprise collaboration system
to get work done. It is less of a meeting place than eRoom. It has good workflow capabilities and
comprehensive document indexing features. The core collaboration functionality lies within iManage
WorkTeam which facilitates the creation of virtual spaces where groups can work on projects.
iManage provides several pre-built templates for creating spaces tailored to each group’s needs.
Within the space, group members can store documents for review, letting iManage’s other main
product, WorkDocs document management system, handle classification and organization. Enterprise
managers can mete out roles to users within the group. Group leaders can create Task Lists, outlining
agenda items and designating specific responsibilities to other users. Communication within the space
occurs within discussion forums, where users can choose to receive automatic notification of any
modifications or replies to postings. Calendar and scheduling functionality is included for meeting
coordination. Users can also synchronize their WorkTeam calendars and Task Lists with Outlook or
Lotus Notes.
Within the iManage’s WorkSite suite, WorkDocs categorizes documents and supports check-in/checkout for secure document editing and keeps detailed histories of document modifications. WorkRoute
presents a good workflow-editing tool. Its GUI interface allows enterprise managers to send key
documents to the appropriate people for review or approval processes. All such processes are saved, so
users can automate this workflow for similar tasks. Managers can view detailed statistics on the
review processes, noting problem steps within the procedure. Employees check their integrated
WorkRoute Inbox for any necessary actions, and once they have handled their own responsibilities,
WorkRoute automatically forwards the document to the next person as specified in the workflow.
Managers can also check the status or progress of any process. These capabilities answer a key concern
of managers – how to monitor and manage workers in a Telework environment.
Using WorkKnowledge users can view and access all content available on WorkSite. Managers can
also filter out important content on WorkDocs or preserve best practices used on a WorkTeam project.
WorkKnowledge can also index and search data gleaned from outside sources such as Lotus Notes
and Web sites. Custom functionality can be added to this collaboration product via the Worksite
Server Development Kit. WorkSite is based on open standards such as Java and XML.
iManage has announced WorkSite MP 3.0, providing tighter overall integration and improved
Telework and Collaboration with Broadband Access
Page 4 of 36
workflow. New features in 3.0 include full J2EE compliance, fully written in Java, XML support, Email is fully integrated into the Workplace core repository, enhanced integration with Microsoft
Word, Excel and Powerpoint including full support for OLE linking and LDAP implementation
enhances security for workgroup access. Not clear if there are any plans for video conferencing and
leveraging video capabilities.
QuickPlace by IBM/Lotus www.lotus.com Lotus QuickPlace is a very comprehensive enterprise
collaboration tool. It also allows users to customize both its look as well as its operation. Lotus
QuickPlace combines a set of collaboration tools into a user-friendly package. It has many advanced
features and customization capability. QuickPlace’s operation centers on the creation of QuickPlaces,
which are essentially virtual spaces for groups to work on projects. A simple browser-based wizard
guides the creation of each QuickPlace. Group leaders can easily select individuals and assign roles for
the project, upon which automatic e-mail notification is sent out to the appropriate parties. QuickPlace
can be integrated with MS Office. Lotus’ own QuickEdit lets users create small documents and
messages, ideal for discussion forums and message boards. QuickPlace offers several methods of
communication within a group. Group chat features and private instant messaging is provided,
however real-time video meeting room capability is not integrated. Group members can check their
colleagues’ online presence since QuickPlace offers LDAP directory integration. Users can also
create discussion threads through QuickEdit. For private group discussions or meetings, users can
create QuickPlace Rooms where only designated group members have access privileges.
QuickPlace’s workflow and task management capabilities are easy to use. Group members can create
task descriptions, assign due dates, designate milestones, and specify tasks as being urgent. Once a task
assignment is made, the item is automatically added to the group calendar and appropriate personal
calendars. Avanced features such as GUI workflow editing are not in QuickPlace, however, the
provided workflow seems to be efficient and user-friendly.
Managers can use Placetypes to preserve best practices from one project for future reuse. These
documents store important project content such as member info, task structure, folder set-up, and
procedures followed in the project. QuickPlace provides most standard editing capabilities, such as
document check-in/check-out, editing permissions, and revision histories.
Lotus QuickPlace 3.0, expands on QuickPlace’s collaborative functionality where .
users can quickly locate and enter all of the team workspaces to which they belong, as well as
search across them. Using a Java/XML API developers can easily embed collaboration in any Web
application. Seamless integration with the Lotus instant messaging product allows real-time
collaboration within a team workspace.
SharePoint Team Services from Microsoft has strong integration with other Microsoft products
and weak integration with non-Microsoft products. Microsoft SharePoint Team Services is a basic
collaboration tool that features an intuitive interface. Many features require separate Microsoft
products (such as MSN Messenger). Although Team Services does offer discussion forums and
document organization capabilities, it noticeably lacks advanced communication features such as
workflow tools and chat software. It also involves manual operation in many areas where automation
would have helped.
The main method of communication provided by Team Services is an asynchronous discussion forum.
SharePoint maintains and organizes all threads of discussion for fast searching. SharePoint also
features an integrated online polling system, in which users can quickly gauge their teams members’
opinions and view the results either graphically or as a Microsoft Excel spreadsheet.
Team Services offers a scheduling tool that allows users to view all team events, personal tasks or
appointments, as well as upcoming events. Users can view these appointments on Team Services’
Calendar or sort and filter events based on a number of metadata characteristics. They can also import
Telework and Collaboration with Broadband Access
Page 5 of 36
appointments from Microsoft Outlook 2002 for immediate inclusion in their list of private events.
Users can set up a document library to hold all content related to a specific issue. They can import any
type of file that they can locate on their desktop, network drives, or on the Web. Administrators can
also specify metadata that must be supplied when submitting a document so content can be
automatically organized. Users can choose to receive automatic notification if a designated document
or library of content is modified. SharePoint lacks many features found in other products on the
market. SharePoint does not provide integrated chat or instant messaging software for when urgent
communication is needed. Furthermore, there are no workflow features to guide shared documents
through the review process. Although SharePoint features Microsoft Outlook integration, it does not
integrate with users’ mail clients.
Open Text Livelink www.opentext.com Open Text’s Livelink is a comprehensive collaboration
solution. The product’s basic functionality addresses many of the key issues in document organization
and team collaboration. It can be integrated with Lotus Notes, Microsoft Outlook, Palm OS.
Livelink’s Virtual Teams combines technology and methodology to create online workspaces for high
performance virtual teams and inter-company networks. During the team creation process, team
roles can also be meted out, simplifying the review process. Employees can coordinate
appointments and meetings using the built-in calendar. Livelink’s MeetingZone add-on provides
further meeting functionality, such as virtual Web-based meeting sessions and instant messaging
capabilities during the live session. However it has no video real-time collaboration software.
Users can also customize every aspect of the software’s operation or augment using one of the many
available add-ons. Through Livelink’s Workflow Designer, workflows can be managed by defining
and modifying user and group participation in workflows, as well as define steps, task profiles
and specific instructions for the process.
The product provides eight levels of user permissions, and the versioning system keeps track of
previous versions of documents and also records when documents were changed, who implemented the
modification, and details on how the document was altered. Users can also choose to be notified of
modifications to any content item. Livelink’s Channel, a part of the core Livelink functionality lets
users broadcast any information or documents to relevant parties within the organizations.
Users can also create detailed workflows using the Designer, specifying review cycles, project goals
and the parties involved in the review process. Corporate executives can designate groups of
employees for work on certain tasks, creating an accompanying workflow using the Designer.
Open Text has provided many avenues for integrating with legacy systems and databases. It is XML
compliant, to integrate disparate applications and data sources. Using one of its many add-ons,
Livelink can interoperate with SAP systems. CORBA is also supported for further ease of integration.
Livelink eSign provides digital signatures. Livelink’s OnTime, a module of Livelink, is a scheduling
and calendaring application which combines full, controlled access to calendar information with good
scheduling functionality.
Although chat and instant messaging features are available in MeetingZone, Livelink does not
provide any bundled presence or chat package. Virtual whiteboards and presence software are
certainly important during group meetings, but real-time conversations and real-time video play an
important role in the everyday workings of a team, so the lack of such modes of communication is a
considerable shortcoming.
Livelink presents a versatile system.
Telework and Collaboration with Broadband Access
Page 6 of 36
ConferenceXP, is a research and development initiative of Microsoft Research Learning Sciences
and Technology group.
ConferenceXP audio and video streams are transported over an RTP network stack. RTP is a peer to
peer network transport that handles stream and message data. Designed for scenarios where latency is
important. RTP is the standard for streaming media in a multicast environment. The ConferenceXP
network transport is expected to dynamically adjusts to handle poor network conditions and low
bandwidth situations. ConferenceXP clients and servers use standard Internet protocols and interfaces
including SOAP, XML, UDDI, WSDL and RTP.
ConferenceXP claims that participants can schedule, join and participate in on-line conferences,
presentations and meetings via a simple point-and-click user interface displayed on their desktop. A
published ConferenceXP API should enable the development of custom user interfaces, and the
integration and use of ConferenceXP with existing conferencing and classroom systems.
The “high speed” of the Internet2 Abilene network, enabled for multicast, is the network being used
for ConferenceXP for high-end collaboration solutions that provide both high quality and low latency
delivery of audio and video.
ConferenceXP is designed to handle full screen video, 30 fps, 250 ms latency. It is designed for
FireWire cameras to enable high quality, efficient video capture, video rate of 256 kb/s 320x240 or
1024 kb/s 640x480 will result in very high quality video resulting in five way conferencing using 2
Mb/s per user. The viewer is designed to have a synchronized display of slides and scribble, comments
and annotations. It is also designed for Integration of slides and whiteboard and enabling annotations.
SECTION II - Results on Telework Experience:
We found market research from Cahners In-Stat/MDR that stated the steady growth of the
remote and mobile (RAM) workforce is continuing and the biggest hurdles to hosted
applications for RAM workers are security and bandwidth constraints. However, the best study
I have seen so far is from AT&T that quantified the improvement due to Telework not only in
productivity gained but also in improving the bottom line from a financial point of view. When
real estate and job retention savings are included, AT&T quotes a savings of more than $100
million every year due to telework and higher bandwidth is a key factor affecting Telework.
The detailed summary below we provide additional details.
In search of demonstrated improvements due to visual collaborations, we found one examples
that, although from a different setting, it relates to the collaboration environment that Telework
Consortium is advocating. At the University of Michigan, Professor Elliot Soloway’s team
http://www.hi-ce.org/ has been studying the application of various tools in K-12 education.
They have found that using palm computing devices and teaching students to beam information
to each other fostered a better collaborative environment then just using PCs asynchronously.
Beaming PDA info forces participants, in real-time to react and work together. A visual
collaboration environment has similar characteristics – team members collaborate better
through interactions, review and creation of new ideas. Dr. Solloway is documenting these
results. A description of some of the tools they have used on PDA’s is in “Palm Computing
Devices are Ready-at-Hand”, Communications of the ACM 2001.
SECTION III - Summary of Technical and Business Status of High speed
Access in the US:
Telework and Collaboration with Broadband Access
Page 7 of 36
We found that there has been much activity in the form of trials and new initiatives on part of
the incumbents and new services by startups that one can learn from. A key consideration in
planning the evolution of a network is the starting point: is there an infrastructure already inplace that can be leveraged or is one designing a brand new network. If it’s a greenfield
scenario, then one would choose the most economical approach that would scale with the
growth of the demand. Next generation technology for building new networks is heading
toward Optical solutions using Ethernet. In contrast, much of the RBOC embedded base is
based on ATM and Broadband PON based on ATM seems to be a key path the RBOCs have
chosen. BPON can offer Ethernet services as well as legacy services.
A key influencing factor in evaluating a migration path to optical Ethernet is the installed base
of broadband capabilities within the metro area(s). Most sites with multiple T1s, T3s and
higher speed interfaces into private line and switched networks are carried over fiber. If fiber is
available it can be extended to the users premises using Passive Optical Networks (PON)
technology. If a SONET ring infrastructure is in place, a packet ring based solution can also
improve the economics of access.
If the infrastructure is not in place, then certainly the approaches used are headed toward a
combination of PON and Packet ring technology. A key issue is the financing of the
investment in high-speed access and traditional service provider approaches or newer
approaches like UTOPIA is using or even more innovative ideas mentioned by Telework
Consortium senior management, needs to be considered and studied.
SECTION IV - AT&T’s Results on Telework In detail:
The Business Value for Teleworking and the need for High Bandwidth Access
Results from AT&T's Dec 2001 annual employee telework research suggest that telework is
being utilized less as an employee perk and more as a mechanism for increasing business
efficiency and continuity. Telework productivity at AT&T has increased roughly 10 percent in
the past year, with employees typically gaining a full hour a day, according to AT&T's eighth
annual telework survey of 1,500 employees. However, the top barrier to telework continues to
be a lack of high-speed data access to the home. Over the last four years, about half of AT&T
managers have worked from home at least once a month, about one-quarter at least once a
week, and about 10 percent in a full-time "virtual office."
AT&T data indicates that teleworking enhances productivity, because teleworkers report being
more productive per unit time and because the teleworker has available the previously nonproductive commute time. When asked about perceived productive work hours (when tasks are
accomplished), office workers reported 6.2 productive hours in an 8 hour day, compared to the
teleworker-reported 7.5 hours in an 8 hour day. Over three-quarters (77%) of all teleworkers
reported higher productivity at home while only 6% reported higher productivity in the office.
Seven-in-ten managers (72%) report being more productive when working from home. Only
5% of managers report higher productivity when working from the office. This increased
productivity was valued by AT&T at $65 million annually. AT&T estimated that they saved
about $25 million per year in real estate through virtual office programs. When real estate and
job retention savings are included, AT&T saves more than $100 million every year due to
telework.
AT&T has several large units within the company that are moving to a full-time virtual office
structure. In 2001, AT&T teleworkers avoided driving 100 million miles, saving 5 million gallons of
gasoline and preventing many thousands of tons of air pollutants. Five of the top six reasons cited by
Telework and Collaboration with Broadband Access
Page 8 of 36
office employees for not working from home relate to the need for speed. Typically, a teleworker who
has a company-paid data line works twice as many days at home per month as one who doesn't. And
those who have high-speed reported gaining about
one
additional productive work hour each day spent working
from home. As a result, hundreds of AT&T employees moved out of traditional offices and are now
working primarily or exclusively from their home offices. AT&T found that 82 percent of teleworkers
said that balancing work and family responsibilities was a significant advantage of telework;

About 70 percent of teleworkers are more satisfied with their current job and their personal and
family lives; and

56 percent of teleworkers who received competing job offers said that they factored telework
into their decision to accept or reject the offer.
Benefits of the Decentralized Organization
AT&T’s experience and data suggests that a decentralized organization - one that is not tied to a single
location, local employees, line-of-sight management cultures, non-networked intellectual capital, or
site-based technology - may be a more efficient, effective, flexible and resilient organization. The
trend toward home-based decentralization is evident when AT&T examined the total population of
those employees who now work from home at least occasionally (see figures below). The total
population of employees who work from home (everyday or 5-19 days per month) increased by
roughly 7%, while the percentage of occasional work-from-homers fell a similar amount. (2000
figures have been adjusted to compensate for the spin-off of AT&T Wireless.)
In this latest research, AT&T teleworkers again reported that they gain about an extra hour of
productive time each day at home, adding up to at least an estimated $65M in business benefit each
year. Separate employee research has shown that full-time virtual office managers receive higher
managerial appraisals than their office-bound peers. Teleworkers get more accomplished not only
because of time saved by not commuting - that is, increased productivity on a per teleworker or per
workday basis - but also because of increased productivity per unit hour. Thirty six percent of those
who stopped teleworking said they were less productive at home because of a lack of technology and
Telework and Collaboration with Broadband Access
Page 9 of 36
27% said they needed face to face communications (which we assume can be addressed by a high
performance video collaboration
environment.
Reasons for AT&T employees stopping telework:
Note that almost one out of five
teleworkers returned to the office
because they were promoted, which is
another indicator of excellent
productivity.
Less productive at home due to lack of technology
36%
Changed job / managers
33%
Lack of face to face communication
27%
Received promotion
19%
Less chance of promotion
15%
Too many distractions at home
2%
Loneliness
1%
Digging deeper into the data, we see
Major Reasons More Employees Don't Telework:
that five out of the top six reasons
Non-Teleworkers Teleworkers
cited by non-teleworkers for lack of
46%
38%
participation involve access speed (i.e. Difficult to download large files
Need to interact with others
48
32
broadband). Work is important to
teleworkers; improving productivity is Slower access to corporate systems
49
33
the second highest reported benefit.
Computer apps don't work well
46
32
Making employees feel trusted is an
Slow access to internet
40
32
under-appreciated advantage; one side Lack of access to broadband
35
33
effect of telework is the management
by results and not appearance. Teleworkers feel that both they and the company save money by the
arrangement, even though many employees use their personal voice lines and other personal
equipment to work from home. The environmental benefits of not commuting are well known - six out
of ten workers say helping the environment is a major advantage of this new work arrangement.
These advantages make teleworkers more loyal to the organization. In AT&T’s study, more than half
(56%) of teleworkers who had received competing job offers factored the ability to work at home into
the decision to stay with the company. If teleworkers were told they could no longer work from home,
one out of three (33%) said they would look for another job within the company - or quit.
There are significant differences between work structures that support frequent telework - even as
often as one or two days a week - and those that support a Virtual Office environment (VO). Almost
anyone can work at home part of the time and the structural elements of the traditional workplace
mostly remain in place (face-to-face meetings, voice and data access, and fixed office space are good
examples, along with other, less visible structures such as security and political systems).
In a virtual office environment, however, the boundaries of location and appearance disappear. It is a
fundamentally different view of an organization and how it works. Meetings and information
exchange can occur anywhere at any time. In a virtual structure, managers must actually understand
the value of each knowledge worker's contribution to the enterprise. The level of trust in a virtual
organization must be much higher than in a frequent telework environment; managers must be
comfortable knowing that the employees they support are working away despite being out of sight,
and employees must feel comfortable that their manager is looking out for them, even when they're not
immediately outside her or his office. Additionally, information technology in the virtual work
environment must evolve so that all required data and functions are available through intranets.
Management and communication channels currently used in the office must have technical equivalents
(video, collaboration systems with shared folders, etc.)
Telework and Collaboration with Broadband Access
Page 10 of 36
AT&T’s view coincides with the Telework Consortium’s view that developing this virtual
infrastructure is a competitive imperative for our society. The economic implications of increasing
productivity by reducing "migration time" are significant, as are the security and business continuity
improvements associated with a decentralized knowledge structure. And these advantages can be
realized in a manner that enhances the quality of life for individual employees, and allows those who
are unable to fully participate in our economy (such as the physically challenged, or those who reside
long distances from economic centers of employment) the ability to contribute based upon knowledge,
not on fate or chance. In addition, clearly the impact on the environment will be very significant.
Lessons from AT&T’s experience
Two different VO (virtual office) initiatives currently being deployed inside AT&T. In these two
organizations alone, more than 500 employees have been moved from dedicated AT&T office space
into virtual offices this year as part of a strategic drive to create a more efficient and effective business
structure. Real estate costs have been reduced by several million dollars, while productivity and job
satisfaction have increased.
Several lessons that illustrate the differences between tactical and strategic telework have emerged
during these initiatives. Three of these lessons are covered below. More will follow in future articles.
Lessons:
1. A critical mass of remote workers in an organization is necessary for management and
communication channels to shift to the network-based equivalents. The first teleworker in an
organization has a very lonely and difficult existence. If a good part of the organization is working
from home, then the cultural barriers drop. Both VO projects took advantage of this principle by
making pilot programs as large as possible. Making the pilot program as large as possible - up to
and including 100% of the proposed work-at-home pool of employees - will uncover technological
and cultural barriers much more quickly because of the diversity of employees, managers and
technologies involved. A large pilot also makes it much easier for teleworkers to feel at ease with
the arrangement, since their peers are also working at home.
2. Need solid technology and policy platforms. It takes a different caliber of infrastructure to support
a 200 person virtual office project than to support two dozen people teleworking occasionally as an
alternative work arrangement. Besides the obvious need for seamless and speedy technology,
remote work must be reflected in the policies and administrative procedures of the company.
3. Recruitment and retention are other important benefits. Because of the enhanced quality of life and
personal freedom which teleworking fosters, firms are better able to retain valued employees, even
when flattening hierarchies so that promotion opportunities are more rare. Among the AT&T
teleworkers who have been offered other jobs, about two-thirds (67%) reported that giving up an
"AT&T telework environment" was a factor in their decision to remain with the company. When
competing for high tech employees, firms are finding that it is the companies with more nontraditional work environments that are the most successful in recruiting the knowledge worker.
OTHER REPORTS: GSA Reports to Congress on IT Barriers to Telework Congress asked GSA
to "identify and develop a plan to resolve technology barrier issues that impede the creation of home
workstations for Federal employees." In response, GSA contracted with the independent research and
consulting firm of Booz Allen and Hamilton to conduct a study of technological barriers to
teleworking and to report on it. The final report on this study was sent to Congress on May 9, 2002.
One result of this study is the following.
The influence of working at home on productivity is important to organizations considering the move
Telework and Collaboration with Broadband Access
Page 11 of 36
to telework. Many managers fear that performance will decline when one is away from the workplace.
Over seventy percent (72.4%) reported that working at home slightly or greatly increased productivity;
19.5% reported that it remained the same; and 6.5% reported that their productivity slightly decreased
since working at home.
Remarks from "The Sky’s The Limit and It’s Not Falling (Yet)" Opening Remarks by Bruce P.
Mehlman Assistant Secretary for Technology Policy, United States Department of Commerce
Delivered January 31, 2002 at the Electronics Industries Alliance Winter Meeting Phoenix, Arizona

ECONOMIC GROWTH: We believe broadband - high-speed, high capacity Internet access and
usage - can further improve U.S. productivity and competitiveness, helping to restore robust
growth to our economy and increasing Americans’ standard of living in the 21st century.

HOMELAND DEFENSE: Broadband can help promote homeland defense. Broadband-enabled
video conferencing will provide important productivity enhancements, allowing the same
economic output despite reduced travel or more distributed organizational structures.
Section V: Details - Status of High Bandwidth Access – the business scenario
Fiber-optic technology, offering virtually unlimited bandwidth potential, is widely considered to be the
ultimate solution to deliver broadband access to the last mile. Today's narrowband telecommunications
networks are characterized by low-speed, service-provisioning delays, and unreliable quality of
service. This limits the ability of workers to be efficient in their jobs. The last mile is the network
space between the carrier's central office (CO) and the subscriber location. This is where bottlenecks
occur to slow the delivery of services.
Incumbent telephone companies responded to Internet access demand by deploying DSL technology. DSL uses the same twisted pair as telephony lines and requires a DSL modem at the
customer premises and a digital subscriber line access multiplexor (DSLAM) in the central office. The
data rate provided by DSL is typically offered in a range of 128 kb/s–1.5 Mb/s. While this is significantly
faster than an analog modem, it is well shy of being considered broadband, in that it cannot support fullservice voice, data, and video. In addition, the distance for DSL coverage is limited to distances less
than 18,000ft. Network operators are now deploying remote DSLAMS closer to subscribers to
aggregate DSL traffic and improve its economics. If the DSLAM is placed within 1000 feet, a flavor of
DSL called VDSL (Very-High-Data-Rate DSL) can be used to provide up to 52 Mbps one-way and 26
Mbps Symmetrical. At 3000 feet, both of the foregoing rates drop in half.
Cable television companies responded to Internet service demand by integrating data services
over their coaxial cable networks, which were originally designed for analog video broadcast.
Telework and Collaboration with Broadband Access
Page 12 of 36
Typically, these hybrid fiber coax (HFC) networks have fiber running between a video headend to one or more hubs and then to multiple fiber nodes. The fiber nodes connect coaxial cable to
multiple distribution amplifiers which feed multiple taps to provide the final drop to the subscribers
The drawback of this architecture is that each shared optical node has less than 36 Mb/s
effective data throughput within a 6 Mhz slot in place of a TV channel using 64 QAM. The
data rate at each node is typically shared by a large number of homes. If the cable provider oversells the bandwidth on a node, low data rates will occur from time to time. Typical downstream
rates range from 500 Kbps to 2500 Kbps. Upstream rates however are about 100 to 200 Kbps. The
upstream data uses frequencies below the TV band which are susceptible to noise and other
services using some of the spectrum. Cable Labs, which is supported by all of the cable companies,
sets the standards which is currently DOCSIS 1.1 (Data Over Cable Service Interface
Specification).The industry, along with Cable Labs, is working on DOCSIS 2.0 with the objective
of achieving symmetrical bandwidth by solving the upstream problem. A likely candidate is to use
Synchronous CDMA. A significant draw back of the DOCSIS standards is the there is no provision
for QoS and SLAs. There is no management of bandwidth allocation. Everyone gets to take part in
the free-for-all except that the provider is able to limit the top rate (around 2500 Kbps) so that no
one user can hog the network.
A new broadband-over-cable technology has been developed by NARAD Networks. It has been in
trial by at least all of the largest cable companies and is starting to be implemented by several of
them. There are also implementations in place in Asia and Europe. A paper giving full details has
been published in the IEEE Communications magazine of August 2002. We have been following
this development for some time and have been in communications with the company by telephone
and a visit to their office in Massachusetts. It is based on Hybrid Fiber Coax (HFC) technology.
DOCSIS uses HFC also but it can be used in a pure coax plant as well. NARAD makes use of the
existing HFC plant that a cable provider has. Since the fiber plant is already in place and well
away from the head end, the coaxial lengths are such that there is much useable coax bandwidth
above the TV frequencies which top out at 865 Mhz. Above that frequency, there is at least 1Ghz
and up to possibly 2Ghz of bandwidth available. The implementation can be done from any node
by replacing the distribution amplifier by a NARAD Network Data Switch (NDS) and replacing the
taps with Subscriber Access Switches (SAS). These devices also provide for passing the legacy TV
services to the subscribers location. The implementation can be done on a market by market basis.
The legacy (TV) service from its headend and Gigabit Ethernet link from a front end rout er are
merged so as to provide both services. The system utilizes 1 Gbps trunks via the coaxial cable and
100 Mbps symmetrical subscriber coax drops. It is based on switched Ethernet
As opposed to DOCSIS services, the NARAD system provides for symmetrical bandwidth at rates
up to 100 Mbps. It also has full QoS and SLA capabilities and software for provisioning and
network management. Each user is guaranteed the bandwidth subscribed. The Cable companies
may be slow to pick up this technology even though the system gives them a leg up compared to
Telcos (CLECs and ILECs). We can only speculate since each cable company may have different
objectives and means. At the present time, even the DOCSIS cable modem services have the lead
over DSL customers by a factor of 2/3 of the “broadband” market vs 1/3 for DSL The total market
currently is about 15.6 million users. (Source: The Companies and Leichtman Group, Inc.). The
market take could be much larger with the NARAD approach
To alleviate bandwidth bottlenecks, optical fibers, and thus optical nodes, are penetrating deeper into
the first mile. The next wave of local access deployment promises to bring fiber to the building
(FTTB) and fiber to the home (FTTH). Unlike previous architectures, where fiber is used as a
feeder to shorten the lengths of copper and coaxial networks, these new deployments use optical
fiber throughout the access network. New optical fiber network architectures are emerging that are
capable of supporting gigabit per second speeds, at costs comparable to DSL and HFC networks. A
Telework and Collaboration with Broadband Access
Page 13 of 36
network infrastructure that allows more bandwidth, quick provisioning of services, and guaranteed
quality of service (QoS) in a cost-effective and efficient manner is now required. Today's access
network, the portion of a public switched network that connects CO equipment to individual
subscribers, is characterized by predominantly twisted-pair copper wiring.
Fiber-optic technology, through local access network architectures such as fiber-to-the-home/building
(FTTH/B), fiber-to-the-cabinet (FTTCab), and fiber-to-the-curb (FTTC) offers a mechanism to enable
sufficient network bandwidth for the delivery of new services and applications. In general, the optical
section of a local access network can either be a point-to-point, ring, or passive point-to-multipoint
architecture.
Fiber-based services such as metro Ethernet are a hot idea, but getting the fiber to the customer isn't
getting much easier. The "fiber glut" in the public network backbone does not extend to the end user.
However, the architecture of metropolitan networks is undergoing a significant shift. With the rise of
carrier hotels and Internet data centers, traffic is being concentrated in new bandwidth "hot spots"
within the metro, these concentration points have become natural targets for large fiber deployments.
Very few office buildings are served by fiber. This is a major missed opportunity for the RBOCs in
meeting customer need and potentially getting ahead of competitors. Carriers have begun collocating
equipment in "carrier hotels" to facilitate their need for many-to-many interconnection. FiberNet
www.fiber.net is an example of the business opportunity that fiber connectivity by itself creates.
(Utopia is another potential example, however it is in the feasibility study stage, prior to floating
bonds).
FiberNet's approach is a focused build out in high density areas. FiberNet Telecom builds fiber
connections among long distance points of presence (POPS), carrier hotels and large office buildings
in New York, Chicago and Los Angeles. FiberNet wholesales to carriers including Broadwing, Qwest
and, most recently, Verizon, as well as to foreign carriers such as Singapore Telecom and Deutsche
Telekom. FiberNet is committed to providing only the fiber infrastructure from the buildings.
FiberNet has access into more than 20 Class A office buildings, including the Chrysler and Seagram
buildings in New York. The company also serves 10 carrier hotels in its three cities, including New
York's 60 Hudson Street, the world's largest such facility.
The incumbent carriers are investing in fiber networks, but with an eye on cost containment as much
as on upgrading customer access. For example, Sprint has focused its metro fiber strategy on securing
dark fiber connections between the IXC's POPS and incumbent local exchange carrier (ILEC) central
offices (COs), Sprint is looking to cut its costs by leasing dark fiber and lighting it up themselves,
rather than leasing lit ILEC facilities. AT&T is looking to technologies such as 38-GHz wireless as a
way of avoiding the cost of running fiber. WorldCom has previously been committed to running fiber
direct to end customers - where it makes sense. That may have been easier for WorldCom, thanks to
acquisitions it made in the late 1990s MFS, Brooks Fiber and MCI all had major metro fiber deployments going back 10 years. WorldCom currently has 50,000 U.S. office buildings and campuses on
fiber in 100+ markets.
Fiber Deployment
The ultimate constraint on the rollout of metro Ethernet service is fiber deployment. The Yankee
group estimated in 2001 that Metro Ethernet providers Market share is 79%, RBOC 19%, IXC 2%.
Connecting a building to a fiber ring is not as expensive as some have estimated ($150,000$400,000). Cogent quotes an average lateral to be "relatively short," about 715 feet. At an average
construction cost of $80 a foot, that would be about $57,000. Cogent’ criteria is that a buildings must
be 100,000 square feet or greater, have 20 or more tenants and be within 1,250 feet of fiber that
Cogent has already secured.
Telework and Collaboration with Broadband Access
Page 14 of 36
The experience with IP public networking over the past couple of years has clearly demonstrated that,
when new technology collides with entrenched carrier business models and infrastructures, the new
technology gets absorbed into the big carriers' networks and business plans in a way that makes sense
for the big carriers. At the end of 2001, the metro Ethernet providers dominated the market that they
had invented (79%). But the RBOCs, without really trying, already own 19 percent of the market,
although for three quarters of 2001, there was only one RBOC with a Gigabit Ethernet offering - most
of their market share came from their transparent LAN services.
The PON Scenario
Even though PON technology has been available for several years, few carriers have rolled out
services over PONs beyond the trial stage. Many of the carriers are planning to use the technology in
the enterprise market, hoping to drive monthly revenues higher with bundled services, but we are not
aware of solid plans for implementation. Several issues have arisen since the standard was set,
including issues around integrating the technology into the carrier operating system, concerns about
fiber costs, and capacity for BPON. Key BPON shortcoming is available Bandwidth. For video
collaboration, we expect symmetrical service will be needed. For symmetrical service, since the total
capacity for BPON to 32 enterprises is limited to 155 Mbps symmetrical (BPON has 622Mbs in
downstream direction and 155Mbs in upstream direction), most of the connections will operate
between a T1 and 4.5 Mbps (equivalent to three T1s). It is only recently that ITU has considered the
recommendation for higher speeds (622Mbs, 1244Mbs, 2488Mbs symmetrical). The original G.983.1
spec has been complemented with G.983.2 and G.983.3, however in general, vendors have yet to
commit to these new BPON (=APON) standards. However, Salira, a start-up from San Jose, Calif.,
has been designing products that go beyond 622 Mbps by leveraging burst-mode optical technology to
deliver Gigabit Ethernet bandwidth and services over PON transport. Salira's solution adds
intelligence at the customer and the service provider end allowing providers flexibility and improved
service-level agreement (SLA) reporting by individual customer port at each node. EPON will offer a
much larger connection to the end user, however, EPON standards are not finalized yet. EPON
systems will have xGbps Capacity. Baseline proposals have been negotiated and agreed to for short
distances and longer distances (12000ft.) standards for Ethernet access. The standards for EPON as part
of IEEE 802.3ah are anticipated by September 2003.
The Incumbents
The established local and long-distance players especially some RBOCs were surprisingly quick to
jump on the Ethernet services bandwagon.
BellSouth announced a Gigabit Ethernet service October 2001, while SBC, which had a point-to-point
GigE-over-fiber offering already in place, said it plans new offerings for 2002. AT&T and Qwest's
IXC division were also planning rollouts of Ethernet service offerings in Tier 1 cities. Qwest is testing
BPON and evaluating EPON. Qwest was planning to build out fiber facilities in 2001 in 25 area
metros. SBC began offering a native Ethernet service (i.e., not run over SONET) called GigaMAN as
early as 1999 in the Ameritech region. The RBOC planned to extend the offering to the rest of SBC's
territory. GigaMAN offers little flexibility on configuration; the only option is 1 Gbps for connecting
two points across a metropolitan area. There's a local-loop charge on either end and mileage charge in
between, with a typical bottom line price of $5,000-$7,000 per month on a 60month contract. SBC is
planning to introduce a multipoint service at 10- and 100 Mbps. Ethernet service is not designed to
have the level of redundancy or reliability that a SONET network has. With GigaMAN, SBC is clearly
out in front of the incumbent pack. Other RBOCs and IXCs are generally building Ethernet services on
top of their existing SONET infrastructure, for added reliability. BellSouth, in 2001 added a Gigabit
Ethernet component to its Native Mode LAN Interconnection (NMLI) offering, its existing transparent
LAN service. The Gigabit service debuted in Atlanta and south Florida, and subsequently expanded
into North Carolina. While SONET is a more expensive infrastructure than native Ethernet, if the
Telework and Collaboration with Broadband Access
Page 15 of 36
investment is already in place then the carrier can leverage it. BellSouth already has the investment in
place, and is planning to start adding next-generation SONET multiservice provisioning platforms
(MSPPs) to its network. Time Warner Telecom has BPON trials targeting small and medium size
enterprises. WorldCom also has planned offering an Ethernet access service (up to 600 Mbps) over its
SONET infrastructure in five cities. Verizon has been evaluating PONs as part of last-mile network
deployment.
On the average, the pricing for optical Ethernet services varies, ranging from $1,570 to $4,040 per
month for 100-Mbps services from Bell South and Telseon respectively. At GigE rates, SBC's tariff is
$8,475 while XO's is $14,000. In 2001, a number of service providers who offered or planned to offer
optical Ethernet services used the following planning numbers (list price) for customer business eases:
$1100/month for 10 Mbps, $2,300/month for 100 Mbps and $11,000/month for 1Gbps, with a
Committed Access Rate (CAR) of $20-$40Mbps. Note that on a $/bit basis, these prices are 30 percent
of the price for 'I'1, 22 percent of a T3 and 20 percent of an OC-3 for nax speeds of 10Mbs, 100Mbs
and 1Gbs respectively..
Case Study: SBC
Most noteworthy development has been SBC's announcement that it would use passive optical
networks (PONS) in the next stage of its Project Pronto initiative to bring broadband to the wider
market. SBC announced it would use a BPON architecture to deliver fiber access to end customers.
SBC’s target market for BPON enabled services is businesses that currently have multiple T1 lines
that are deployed using repeaters. Thus, the carrier's focus is on saving money within its infrastructure
(however, they are also looking at Ethernet services). Shifting repeatered-T1 customers over to BPON
also would improve the range and service quality of SBC's DSL offerings, since T1 signals can interfere with DSL. With Capital Expenditures Capped, the rate of deployment is not expected to be of a
very large scale.
SBC Broadband PON (BPON) Architecture
One fiber connection serves many homes or businesses and cascaded couplers lower cost
Telework and Collaboration with Broadband Access
Page 16 of 36
Companies such as BellSouth have deployed PONs in their fiber-to-the-home (FTTH) trials. While
PONs brought down the price of fiber access to homes, the capital expense is still $350–$450 per
home plus the cost of the fiber that can range in cost from $40–$100 per foot depending on location.
This is significantly more than the upgrade costs for DSL installation, causing many carriers to take a
wait-and-see attitude toward widespread FTTH deployments. To improve the return on investment for
their fiber rollout, some carriers and service providers are testing their PON technology within the
enterprise market.
The Startups. Leaders: Cogent, Yipes, OnFiber. Challengers: IntelliSpace, Looking Glass.
Yipes is a national provider of Ethernet-based IP services to enterprise customers. Yipes' architecture
uses Ethernet to carry IP packets over optical fiber. Yipes Communications filed a voluntary petition
for relief under Chapter 11 in late March of 2002. In just over three months, Yipes Enterprise
Services was formed, formulated a revised business plan for the Yipes services, raised equity
financing and consummated the acquisition of Yipes Communications assets. Under the new business
Telework and Collaboration with Broadband Access
Page 17 of 36
plan, Yipes Enterprise Services will continue to serve customers and expand services in 10 markets
including San Francisco, San Diego, Seattle, Chicago, New York, Philadelphia, Denver, Dallas,
Houston and Washington D.C. Customers in remaining markets are planned to transition to Expedient Holdings, Inc., a national service provider. Yipes is currently operating in 20 Tier 1 MSAs.
has targeted large enterprises with LAN interconnection and Internet access services, and stressed its
ability to offer services up to 1 Gbps in increments as granular as 1 Mbps. Yipes pricing was 40-60
percent less than the incumbents' rates for the same bandwidth level. Yipes had 700 customers as of
the Chapter 11 filing: One of Yipes differentiators was the ability to provide bandwidth on demand, in
very fine (1M) increments. However, a key issue is the coverage area.
Yipes has a head start on similar EtherLECs in developing its suite of IP services that will be bundled
for clients. Services include:
Yipes MAN. Connects multiple client locations using native LAN speeds of 1 Mbps to 1 Gbps.
Yipes WAN. Allows locations in different cities to connect with an IPVPN using IPSec protection
at speeds of 1 Mbps to 100 Mbps.
Yipes Net. Gives customers Internet access at speeds of 1 Mbps to 1 Gbps through a Yipes
peering partner. Multiple partnering agreements with Tier 1 ISPs. Access services monitored by
Yipes-maintained NOC.
Yipes Wall. A security firewall solution.
Yipes Web. Collocation, managed hosting services, managed servers, and professional services.
Cogent provides Metro Ethernet service with revolutionary pricing levels: A full 100 Mbps for a flat
$1,000 a month. But Cogent only offers Internet access, it does not sell increments of bandwidth.
Cogent has a nationwide fiber backbone, so a customer could run its own VPN over a Cogent Internet
access service, but typical service is to a single location.
Other significant retail Ethernet providers have included FiberCity, GiantLoop, Intellispace and XO
Communications. Time Warner Telecom, a business-oriented CLEC, is offering Ethernet services
in all 44 of its markets as well.
Ethernet startups have gone to great lengths to provide redundancy, alternate routing and other means
of providing greater reliability in their native-Ethernet networks. The metro Ethernet providers have
always maintained that their recovery metrics were adequate, if not top-of-the-line SONET level.
Yipes, for example, uses redundant switch/routers on a fiber ring, and can switch over in a matter of
seconds vs. SONET's 50millisecond guarantee. Thus, SONET is designed to replace private lines,
whereas Ethernet on Fiber addresses IP Service (where applications can handle outage of a few
seconds. SBC’s GigaMAN is a dedicated, point-to-point connection, and SBC offers service level
agreements (SLAs).
OnFiber Communications: Offers DWDM, Ethernet, and SONET-based services to carriers, service
providers, and enterprises. Acquired two competitors, Sphera Communications and Telseon
Communications, to expand its network reach and customer base. Telseon has taken a different track
than the other providers and was looking to be the backbone of the ISPs, ASPs, CDSPs, and SANs that
need to move large amounts of data around the metro area. The company is also serving Web
enterprises that need to connect to the facilities of their service providers. Telseon is servicing 20 Tier
1 MSAs in the United States. The company competes with the RBOCs while complementing the
metro extensions of the greenfield IXCs. Telseon has agreements with 360networks (bankrupt) and
Level 3 to provide their customers with connections to metro facilities that are outside the current
reach of the IXC’s networks. This will drive traffic on Telseon’s network as well as provide an endto-end IP solution for a service provider customer needing transport in multiple metros. In addition,
Telseon announced that it had inked a partnership agreement with Williams Communications to share
Telework and Collaboration with Broadband Access
Page 18 of 36
access to network points of presence (POP). Under the agreement, Telseon will have access to
Williams' POPs and long-haul backbone for its new managed wavelength service, adding 100 POPs to
Telseon's network reach. In exchange, Williams will have access to Telseon's current 125 connected
POPs to deliver its data services. Telseon has also signed multiple agreements with established players
to extend its reach into hundreds of end-user buildings.
Giant Loop offers a suite of managed services to Global 2000 companies within Tier 1 MSAs to
address both storage and data networking. Giant Loop has a presence in five cities and planned to
expand into international financial capitals. Called Enterprise Optical Networking (EON), Giant Loop's
offering provides point-to-point, point-to-multipoint, and multipoint-to-multipoint multi-gigabit
connections. EON will support IP, Gigabit Ethernet, ATM, Fiber Connection (FICON), Enterprise
Systems Connectivition (ESCON), and Fibre Channel. Giant Loop will also offer a range of
professional services to provide a complete data and storage protocol service.
Intellispace has been serving small and medium businesses with Ethernet services for more than six
years. It has maintained a steady growth in customer numbers with Internet access, security, storage
and hosting services. Targeting the Tier 1 MTU market for its rollout of Ethernet-based services, has
wired over 100 million square feet of office space.
Enterprises face an array of decisions as they plan their migration to fiber-based Ethernet in the
campus, MAN and WAN. CIBC, a leading financial institution, is rolling out a managed optical
Ethernet network from Bell Nexxia, starting with major sites around Toronto. Kaiser Permanente, has
issued an RFI for a nationwide optical Ethernet-based solution spanning almost 500 sites. Nortel
Networks has had a project converting its network architecture from ATM to native Ethernet among
its major sites across North The rationale for these activities is straightforward: simpler, faster and
more reliable networking, opportunities for rethinking server and storage distribution, and increased
knowledge-worker productivity.
Metromedia Fiber Network (MFN): Founded in 1993, MFN was one of the first to competitively
deploy fiber networks in the metro for carriers. A large debt load forced the provider into Chapter 11
protection in May 2002 and an SEC investigation was announced in June. MFN is currently working
on a restructuring plan.
Summary of Metro Ethernet Case Study by METRO ETHERNET FORUM:
Another study that provides insight was carried out by the Metro Ethernet Forum. The study
analyzed the potential savings over a three-year period realized by metro Ethernet data services over
today's traditional options. For the purposes of this case study, only savings resulting from lower
monthly recurring charges (MRCs) were calculated. It should be noted that the study did not quantify
the additional savings generated by engineering and operational support efficiencies and by reduced
CPE interface and platform capital expenditures. Although not addressed, these can be substantial
savings that can further add to those afforded by reduced MRCs.
To quantify MRC savings, the case study modeled a representative metro area profile for a large
nationwide enterprise. The modeled metro area consisted of one large site with 500+ employees and
four medium Sites with 100-499 employees. It was assumed that these sites require connectivity for two
metro data applications 1) Dedicated Internet Access, 2) Private data networks. Bandwidth demands
for each of these applications were assumed for the three year period and applied to three metro models.
Telework and Collaboration with Broadband Access
Page 19 of 36
Table 1: Assumptions for Three-Year Bandwidth Demand Per Application
Bandwidth Demand
(Mbps)
Dedicated Internet Access
Private Data
Dedicated Internet Access
Private Data
Year 1
3.5
3.0
10.0
25.0
Year 2
4.6
3.6
13.0
30.0
Year 3
5.9
4.3
17.0
36.0
Table 2: Assumption for Metro Data Service Models
Model 1
Model 2
Model 3
Traditional Metro
Traditional Metro
Metro Ethernet
Private Line
Frame Relay
Dedicated Internet Access Dedicated Internet Access Ethernet Internet Access
Local Private Line
Local Frame Relay
Metro Ethernet Private Line
Traditional and metro Ethernet service pricing was then applied to calculate total annual costs for each
of the three models. Traditional service pricing was derived from publicly available government tariffs
and other published sources and is representative of the industry average for U.S. IXCs and ILECs. In
the absence of widely published list pricing for metro Ethernet services, the case study
leveraged pricing information gathered from interviews with the major U.S. IXCs, ILECs, and
CLECs offering Metro Ethernet Private Line and Ethernet Internet Access services. The metro
Ethernet service pricing used is a conservative representative of the middle tier between the lowest and
highest priced players.
Telework and Collaboration with Broadband Access
Page 20 of 36
The following additional assumptions/qualifications were are also made to calculate recurring
metro data service costs for all models:
 Service Pricing Assumptions/Qualifications:
Service pricing is based solely on monthly recurring charges (MRCs) and does not include
Non-recurring charges (NRCs). All service pricing (traditional and Ethernet-based):
 Decreases 10% annually
 Includes local access charges
 Is based on a minimum 2-year contract
Distance-sensitive local private line pricing is based on the following assumptions:
 3 mile average local access distance
 20 mile average interoffice distance

Bandwidth Growth Assumptions:
 Forecasted bandwidth is for data transport only
 Bandwidth demand for each application increases annually at the following rates:
 Internet Access: 30%
 Private Data: 20%
Free Space Optics
Another technology developed for faster and cheaper deployment of optical access to enterprises is
free space optics (FSO), an option that carriers will use along with PON as a low-cost alternative to
point-to-point fiber deployment. With PONs and FSO, carriers will be able to reach a greater number
of potential customers with optical access. This line-of-sight technology uses lasers to transmit data
without fiber; instead, the data is transmitted through the air. The technology has been around for
years in campus environments, connecting short distances between buildings. Recently, vendors have
improved their systems to increase distance capabilities, reliability, and capacity of the systems, with
ATM-based FSO operating up to 622-Mbps and 1-Gbps systems using Ethernet.
FSO has several advantages as well as disadvantages compared with fiber connections.
These advantages include lower deployment cost, Faster provisioning, Portability. One example of
a temporary deployment occurred in New York City after September 11, when the technology was
used to connect investment bankers with broadband access while they were relocated to temporary
offices.
Telework and Collaboration with Broadband Access
Page 21 of 36
Disadvantages for FSO versus fiber include: Reliability. FSO systems are affected by weather,
particularly fog and snow. Radio backup can help with reliability, since the radios can operate without
a problem in the fog. Distance limitations. FSO deployments with carrier-class reliability are limited
to less than a mile, depending on the climate within a given metro. Direct Line of Sight. This will
limit carrier FSO deployment to locations in the central business districts of metro areas.
Telework and Collaboration with Broadband Access
Page 22 of 36
Section VI: Additional Technical discussion of Broadband Access
This section focuses on providing more technical detail. To be self consistent, we provide some of
the detail described in section II, the business focused section.
In metropolitan areas, where there is a high concentration of business customers, the access network
often includes high-capacity synchronous optical network (SONET) rings, optical T3 lines, and
copper-based T1s. Digital subscriber line technology (xDSL) and cable modems offer a more
affordable solution for data, but they are not widely available and difficult and time-consuming to
provision. In addition, bandwidth is limited by distance and by the quality of existing wiring and
carriers are aggregating customer traffic via shared access (Cable) or stat muxing at DSLAMs
(xDSL) .
SONET/SDH networks use complex ring topologies and are designed as contiguous and redundant
point-to-point solutions with expensive interconnection nodes known as Add/Drop Multiplexers
(ADMs). At an ADM, an optical-to-electrical-to-optical conversion is done and the overall network
is optimized for either long-haul or metropolitan applications. To reduce costs and expand the
market for fiber access, carriers and vendors have worked together to develop Passive Optical
Networks (PONs) to address the last mile of the communications infrastructure between the service
provider’s CO, head end, or point of presence (POP) and business or residential customer locations.
PON architecture eliminates the active network components between the customer premises and the
central office (CO), cable headend, or carrier point-of-presence (POP) required in a traditional fiber
network, thereby reducing initial equipment costs and maintenance expenses. Active components
between the CO and the premises require power and must be hardened against weather and
temperature extremes. By replacing these components with a passive optical splitter, carriers can
reduce the costs of deploying optical access while increasing available bandwidth and services to
their customers.
The PON architecture was designed to bring fiber access to a larger share of the residential and
enterprise markets. Brought together by BT under the Full-Service Access Network (FSAN)
Coalition, NTT, BellSouth, and 17 other large carriers began developing PON standards in 1995
based on asynchronous transfer mode (ATM) standard to transport data, video, and Ethernet
services. The International Telecommunication Union (ITU) has accepted this format for ATMover-PON (APON), which has been deployed in limited fashion by BellSouth, SBC, and enterprisefocused units of cable operators. To avoid confusion and the impression that APON can only carry
ATM traffic, the FSAN Coalition has changed the name of the APON standard to broadband-overPON (BPON). Vendors are currently working on Ethernet-over-PON (EPON), and new services
have been built on EPON as well. Vendors are lowering the costs of PON technology using EPON
while increasing PON bandwidth capacities.
Even as the access network remains at a relative standstill, bandwidth has increased dramatically on
longhaul networks through the use of wavelength division multiplexing (WDM) and other new
technologies. WDM technology has begun to penetrate metropolitan-area networks (MAN), boosting
their capacity dramatically. At the same time, enterprise local-area networks (LAN) have moved
from 10 Mbps to 100 Mbps, and soon many LANs will be upgraded to gigabit Ethernet speeds. The
result is a growing gulf between the capacity of metro networks on one side and end-user needs on
the other, with the last-mile bottleneck in between. PONs aim to break the last-mile bandwidth
Telework and Collaboration with Broadband Access
Page 23 of 36
bottleneck by targeting the sweet spot between T1s and OC–3s that other access network
technologies do not adequately address.
In our investigation, we will focus on the next generation, lower cost access technology. Our aim is
to investigate the technology that can provide the bandwidth needs for true collaboration, investigate
collaboration tools and demonstrate the use via laboratory, field trials and customer experience.
BANDWIDTH ESTIMATES FOR SERVICES
Service
Collaborative remote studio, video editing
Full-motion videoconferencing or streaming
Required Bandwidth
6 - 45 Mbps
3 - 6 Mbps (MPEG)
Lower-quality videoconferencing
384Kbps – 1.5Mbs
Encoding technique Bit rate
H.26
1
M-JPEG
MPEG-1
MPEG-2 (H.262)
Resolution
64 Kbps-2
Mbps
3-8 Mbps
15-25
60-100
Mbps Mbps
1.2-3 Mbps
5-10 Mbps
20-40
Mbps
1-2 Mbps
4-5 Mbps
8-10 Mbps
20-30
Mbps
Broadcast standard
176x144
352x288
QCIF (conference)
CIF (VHS quality)
352x288
720x486
1920x1080
352x288
720x486
1920x1080
352x288
720x486
960x576
1920x1080
CIF (VHS quality)
CCIR601 (PAL)
HDTV
CIF (VHS quality)
CCIR601 (PAL)
HDTV
CIF(VHS quality)
CCIR601
EDT
(Pal)
HDTV
What is PON?
Passive Optical Networks (PONs) are low cost optical Fiber-to-the Building/Curb/Home (FTTb,
FTTc, FTTh or collectively referred to as FTTx) solutions. PON is a point-to-multipoint optical
network that allows service providers to minimize the need for fiber in the outside portion of the
network (e.g., outside plant-OSP) to interconnect buildings or homes. As a result, PONs cost a
fraction of what it takes to deploy new point-to-point fiber or to rehabilitate existing SONET rings.
Telework and Collaboration with Broadband Access
Page 24 of 36
In the Figure below a) shows a point-to-point (P2P) topology, with dedicated fiber runs from the local
exchange to each end-user subscriber. While this is a simple architecture, it is cost prohibitive due
to the fact that it requires significant outside plant fiber deployment as well as connector termination
space in the local exchange. Considering N subscribers at an average distance L km from the central
office, a P2P design requires 2N transceivers and N * L total fiber length (assuming single fiber is
used for bidirectional transmission). In b) we show a remote concentrator close to the neighborhood.
This reduces fiber consumption to only L km (assuming negligible distance between the switch and
customers), but actually increases the number of transceivers to 2N + 2. Curb-switched architecture
requires electrical power as well as backup power at the curb unit. Currently, one of the highest costs
for local exchange carriers is providing and maintaining electrical power in the local loop. Therefore, it
is logical to replace the hardened active curb-side switch with an inexpensive Passive Optical Solution
as shown in c). PONs minimize the amount of optical transceivers, central office terminations, and
fiber deployment. PON is a point-to-multipoint optical network with no active elements in the signals’
path from source to destination. The only interior elements used in a PON are passive optical components,
such as optical fiber, splices, and splitters. Access networks based on single fiber PON only require N+1
transceivers and L km of fiber.
A typical PON solution requires two types of interconnected devices: equipment located in a service
provider's local exchange (or headend) known as the Optical Line Terminal (OLT); and multiple
Optical Network Units (ONUs) geographically distributed in buildings, on curbs, on utility poles or
on the sides of homes. A high bandwidth optical signal is sent on a single optical fiber line, and then
optically split to several ONUs. An ONU receives, and in turn transmits, an independent wavelength,
on the same fiber strand, and provides end users with dynamically allocated bandwidth for voice,
data and video services. On the upstream path, traffic from the ONUs is aggregated back to the OLT
using network topologies such as tree, bus or fault-tolerant rings.
Thus, the PON system uses a double-star architecture. The first star is at the OLT, where the widearea network interface to services is logically split and switched to the PON interface. The second
star occurs at the splitter where information is passively split and delivered to each ONT. The OLT
Telework and Collaboration with Broadband Access
Page 25 of 36
is typically located in the carrier's CO. The OLT is the interface point between the access system
and service points within the carrier's network. When data content from the network reaches the
OLT, it is actively switched to the passive splitter. The OLT behaves like an ATM edge switch with
PON interfaces on the subscriber side and ATM–synchronous optical network (SONET) interfaces
on the network side.
The active network components are the Optical Line Terminal (OLT) at the CO and the Optical
Network Unit (ONU) at the premises. For a BPON network, the SONET add/drop multiplexer
(ADM) is replaced by the OLT at the CO and the ONU at the customer premises. With an EPON
deployment, the OLT replaces the ATM switch and SONET ADM in the CO, while the ONU
replaces the SONET ADM and router at the customer premises. This streamlines the architecture,
and reduces costs and maintenance.
A design issue for PON technology is its lack of built-in redundancy when configured in its tree
architecture. (In SONET for example a secondary ring provides almost immediate (50ms)
reconfiguration for reliability). This is not an issue in cases where a PON is replacing a traditional
point to point private lines, where one has to pay for a second line to obtain redundancy.
Furthermore, given that PONs use passive components, the failure rate should be extremely low.
Nevertheless, in order for PONs to have the redundancy offered by a point-to-point fiber pair, which
can ensure availability through optical redundancy or route-diverse connections, several ONUs can
be installed on the premises or the network must be deployed in the architectures shown in the
figure below. The choice among these architectures will be based on cost of deployment and level
of availability desired.
The two primary types of PON technology are asynchronous transfer mode PONs (APONs) and
Ethernet PONs (EPONs).
Telework and Collaboration with Broadband Access
Page 26 of 36
BPON (aka APON)
BPONs were developed in the 1990s through the work of the full-service access network (FSAN)
initiative. FSAN was a group of 20 large carriers that worked with their strategic equipment
suppliers to agree upon a common broadband access system for the provisioning of both
broadband and narrowband services. British Telecom organized the FSAN Coalition in 1995 to
develop standards for designing the cheapest, fastest way to extend emerging high-speed services,
such as Internet protocol (IP) data, video, and 10/100 Ethernet, over fiber to residential and
business customers worldwide.
At that time the two logical choices for protocol and physical plant were ATM and PON: ATM
because it was thought to be suited for multiple protocols, PON because it is the most economical
broadband optical solution. The APON format used by FSAN was accepted as an International
Telecommunications Union (ITU) standard (ITU–T Rec. G.983). The ITU standard focused
primarily on residential applications and in its initial version did not include provisions for
delivering video services over the PON. Subsequently, a number of start-up vendors introduced
APON–compliant systems that focused exclusively on the business market. The ITU Spec G.983.1
provides for alternative architectures leading to higher availability.
SBC ATM PON Characteristics

622Mbs (or 155Mbs) downstream and 155 Mbps upstream on a single feeder fiber
 WDM upgrades available (Broadband PON)
 two wavelengths are used—1550 nm for the downstream and 1310 nm for the upstream.
 ATM-based




TDM downstream; TDMA upstream
does not use SONET frame
uses ATM CBR circuit emulation for T1 transport
uses ATM UBR, VBR for transparent LAN services
 Maximum reach 20 km
 SBC APON will support SVCs, VoATM, VoIP
The PON specification allows for up to 64 locations to be served. Thus, a typical BPON system can
furnish up to 64 customer locations on a single, shared strand of fiber running at 155 Mbps. Most,
however, are expected to initially utilize 32 locations in the distribution and drop portion of the
network.
EPON
Ethernet passive optical networks (EPON) are an emerging access network technology that provides
a low-cost method of deploying optical access lines between a carrier’s central office (CO) and a
customer site. This represents a that can ultimately lead to widespread adoption of a new optical IP
Ethernet architecture that combines the best of fiber optics and Ethernet technologies. This
architecture is aimed at delivering bundled data, video, and voice services over a single platform.
EPONs offer the highest bandwidth to customers of any PON system today. Downstream traffic
Telework and Collaboration with Broadband Access
Page 27 of 36
rates of 1 Gbps in native IP have already been achieved, and return traffic from up to 64 ONUs can
travel in excess of 800 Mbps.
The development of EPONs has been spearheaded by one or two visionary startups that feel that the
BPON standard is an inappropriate solution for the local loop because of its insufficient bandwidth,
its complexity, and its expense. Also, as the move to fast Ethernet, gigabit Ethernet, and now 10gigabit Ethernet picks up steam, these start-ups believe that EPONs will eliminate the need for
conversion in the wide-area network (WAN)/LAN connection between ATM and IP protocols.
EPON vendors are focusing initially on developing fiber-to-the-business (FTTB) and fiber-to-thecurb (FTTC) solutions, with the long-term objective of realizing a full-service fiber-to-the-home
(FTTH) solution for delivering data, video, and voice over a single platform. EPONs offer higher
bandwidth, lower costs, and broader service capabilities than BPON.
In November 2000, a group of Ethernet vendors kicked off their own standardization effort, through
the formation of the Ethernet in the First Mile (EFM) IEEE study group. The new study group aims
to develop a standard that will apply the proven and widely used Ethernet networking protocol to the
access market. Sixty-nine companies, including 3Com, Alloptic, Cisco Systems, Intel, MCI
WorldCom, and World Wide Packets, have participated in the group.
IEEE P802.3ah STATUS
The standards work for Ethernet in the local subscriber access network is being done in the IEEE
P802.3ah Ethernet in the First Mile (EFM) Task Force www.ieee802.org/3/efm ,
www.ieee802.org/3/efm/public. The IEEE P802.3ah EFM Task Force is bringing Ethernet to the local
subscriber loop, focusing on both residential and business access networks. Requirements of local
exchange carriers are vastly different than those of enterprise managers for which Ethernet was
designed. In order to “evolve” Ethernet for local subscriber networks, P802.3ah is focused on four primary
standards definitions:
1. Ethernet over copper
2. Ethernet over Point to Point fiber
3. Ethernet over P to Multipoint fiber (EPON)
4. Operation, administration, and maintenance (OAM)
Thus, the EFM Task Force is focused on both copper and fiber standards, optimized for the first
mile and glued together by a common OA&M system. This is a particularly strong vision, since it allows
a local network operator a choice of Ethernet flavors using a common hardware and management
platform. In each of these subject areas, new physical layer specifications are being discussed to meet
the requirements of service providers while preserving the integrity of Ethernet. The Ethernet over
P2MP track is focusing on the lower layers of an EPON network. This involves a PHY
specification, with possibly minimal modifications to the 802.3 MAC. This emerging protocol uses
MAC control messaging (similar to the Ethernet PAUSE message) to coordinate multipoint-to-point upstream Ethernet frame traffic. Baseline proposals have been submitted and are being negotiated.
For example, recently an agreement was reached on standards for short distances (750ft) and longer
distances (12000ft.). Long haul service will be able to aggregate 2MBs BW to reach the higher BW
available at short distances. The standards for EFM are anticipated by September 2003.
The key difference between EPONs and BPONs is that in EPONs, data is transmitted in variablelength packets of up to 1,518 bytes according to the IEEE 802.3 protocol for Ethernet, whereas in
Telework and Collaboration with Broadband Access
Page 28 of 36
APONs, data is transmitted in fixedlength 53-byte cells (with 48-byte payload and five-byte
overhead), as specified by the ATM protocol. This format means it is inefficient for BPONs to carry
traffic formatted according to IP. For BPON to carry IP traffic, the packets must be broken into 48byte segments with a 5-byte header attached to each one.
Economic Case for Ethernet PONs
The economic case for EPONs is simple: fiber is the most effective medium for transporting data,
video, and voice traffic, and it offers virtually unlimited
bandwidth. But the cost of running fiber “point-to-point” from every customer
location all the way to the CO, installing active electronics at both ends of each fiber, and managing
all of the fiber connections at the CO is prohibitive. EPONs address the shortcomings of point-topoint fiber solutions by using a point-to-multipoint topology instead of point-to-point in the outside
plant; by eliminating active electronic components, such as regenerators, amplifiers, and lasers, from
the outside plant; and by reducing the number of lasers needed at the CO.
Comparison of Point-to-Point Fiber Access and EPONs
Point-to-Point Fiber
Access
EPON
Point-to-Point Architecture
Point-to-Multipoint Architecture
Active electronic
components are required at
the end of each fiber and in
the outside plant.
Eliminates active electronic components, such as
regenerators and amplifiers, from the outside plant and
replaces them with less-expensive passive optical
couplers that are simpler, easier
to maintain, and longer lived than active components
Each subscriber requires a
Conserves fiber and port space in the CO by passively
separate fiber port in the CO. coupling traffic from up to 64 optical network units
(ONU) onto a single fiber that runs from a
neighborhood demarcation point back to the service
provider’s CO, head end, or POP
Expensive active electronic Cost of expensive active electronic components and
components are dedicated to lasers in the optical line terminal (OLT) is shared over
each subscriber
many subscribers
Unlike point-to-point fiber-optic technology, which is optimized for metro and longhaul
applications, EPONs are tailor-made to address the unique demands of the access network. Because
they are simpler, more efficient, and less expensive than alternative access solutions, EPONs finally
make it cost-effective for service providers to extend fiber into the last mile and to reap all the
rewards of a very efficient, highly scalable, low-maintenance, end-to-end fiber-optic network.
The key advantage of an EPON is that it allows carriers to eliminate complex and expensive
asynchronous transfer mode (ATM) and SONET elements and to simplify their networks
dramatically (if these technologies have not been deployed before). In a streamlined EPON
Telework and Collaboration with Broadband Access
Page 29 of 36
architecture, an ONU replaces the SONET ADM and router at the customer premises, and an
OLT replaces the SONET ADM and ATM switch at the CO. This architecture has lower up
front capital equipment and ongoing operational costs relative to SONET and ATM. EPON is
easier to deploy than SONET/ATM because it requires less complex hardware and no outside
plant electronics. This architecture can deliver bandwidth in scalable increments from 1 to 100
Mbps and up to xGbps and value-added services, such as managed firewalls, voice traffic
support, VPNs, and Internet access.
The passive elements of an EPON are located in the optical distribution network (also known as the
outside plant) and include single-mode fiber-optic cable, passive optical splitters/couplers,
connectors, and splices. Active NEs, such as the OLT and multiple ONUs, are located at the end
points of the PON. Optical signals traveling across the PON are either split onto multiple fibers or
combined onto a single fiber by optical splitters/couplers, depending on whether the light is traveling
up or down the PON. The PON is typically deployed in a single-fiber, point-to-multipoint, tree-andbranch configuration for residential applications. The PON may also be deployed in a protected ring
architecture for business applications or in a bus architecture for campus environments and multipletenant units (MTU).
The interface between the customer’s data, video, and telephony networks and the PON is via the
ONU. The primary function of the ONU is to interface the traffic in an optical format and map it to
and from the customer’s desired format. A unique feature of EPONs is that, in addition to
terminating and converting the optical signal, the ONUs provide Layer-2 and -3 switching
functionality, which allows internal routing of enterprise traffic at the ONU. EPONs are also well
suited to delivering video services in either analog CATV format, using a third wavelength, or IP
video.
Because an ONU is located at every customer location in FTTB and FTTH applications and the costs
are not shared over multiple subscribers, the design and cost of the ONU is a key factor in the
acceptance and deployment of EPON systems. Typically, the ONUs account for more than 70
percent of the system cost in FTTB deployments, and in FTTH deployments they account for
approximately 80 percent.
Key features and functions of the ONU include the following: Customer interfaces for POTS, T1,
DS–3, 10/100BASE-T, IP multicast, and dedicated wavelength services. Layer-2 and -3 switching
and routing capabilities. Provisioning of data in 64 kbps increments up to 1 Gbps. Note that Standard
Ethernet interfaces eliminate the need for additional DSL or cable modems
EPONs offer many cost and performance advantages that enable service providers to deliver
revenue-generating services over a highly economical platform. However, a key technical challenge
for EPON vendors lies in enhancing Ethernet’s capabilities to ensure that real-time voice and IP
video services can be delivered over a single platform with the same QoS and ease of management
as ATM or SONET.
EPON vendors are attacking this problem from several angles. The first is to implement methods,
such as differentiated services (DiffServ) and 802.1p, which prioritize traffic for different levels of
service. One such technique, TOS Field, provides eight layers of prioritization to make sure that the
Telework and Collaboration with Broadband Access
Page 30 of 36
packets go through in order of importance. Another technique, called bandwidth reserve, provides an
open highway with guaranteed latency for POTS traffic so that it does not have to contend with data.
Comparison of SONET, and EPON Service Objectives and Solutions
Objective
ATM/SONET Solution
Ethernet PON Solution
Real-time
services
ATM service architecture and
connection-oriented design
ensure the reliability and
quality needed for real-time
service.
A routing/switching engine that offers native
IP/Ethernet classification with advanced admission
control, bandwidth guarantees, traffic shaping, and
network resource management that extends
significantly beyond the Ethernet solutions found
in traditional enterprise LANs
Statistical
Traffic shaping and network
multiplexing resource management allocates
bandwidth fairly between users of
non–real-time services Dynamic
bandwidth allocation
implementation needed.
Traffic-management functionality across the
internal architecture and the external interface with
the MAN EMS provides coherent policy based
traffic management across OLTs and ONUs. IP
traffic flow is inherently bandwidth conserving
(statistical multiplexing).
Multiservice These characteristics work
delivery
together to ensure that fairness
is maintained among different
services coexisting on a common
network.
Service priorities and SLAs assure that
network resources are always available for
a customer-specific service. Gives service
provider control of “walled-garden” services, such
as CATV and interactive IP video.
Management A systematic provisioning
capabilities framework and advanced
management functionality
enhance the operational tools
available to manage the network.
Integrating EMS with service providers’
OSSs emulates the benefits of connection
oriented networks and facilitates end-to
end provisioning, deployment, and
management of IP services.
Protection
Counter-rotating ring architecture provides
protection switching in sub 50 ms intervals.
Bidirectional line-switched ring
(BLSR) and unidirectional
path-switched ring (UPSR)
provide full system redundancy
and restoration.
ILECs realize that T1 services are their “bread and butter” in the business market. However, T1 lines
can be expensive to maintain and provision, particularly where distance limitations require the use of
repeaters. Today, most T1s are delivered over copper wiring, but service providers have already
recognized that fiber is more cost-effective when demand at a business location exceeds four T1
lines.
EPONs provide the perfect solution for service providers that want to consolidate multiple T1s on a
single cost-effective fiber. By utilizing a PON, service providers eliminate the need for outside plant
Telework and Collaboration with Broadband Access
Page 31 of 36
electronics, such as repeaters. As a result, the expense required to maintain T1 circuits can be
reduced dramatically. In many cases, savings of up to 40 percent on maintenance can be achieved by
replacing repeatered T1 circuits with fiber-based T1s.
EPONs provide the most cost-effective means for ILECs, CLECs, and MSOs to roll out new,
higher-margin fast Ethernet and gigabit Ethernet services to customers. Data rates are scalable from
1 Mbps to 1 Gbps, and new equipment can be installed incrementally as service needs grow, which
conserves valuable capital resources. In an analysis of the MSO market as claimed by Alloptics, an
FTTB application delivering 10/100BASE-T and T1 circuits yielded a one-month payback
(assuming a ratio of 70 percent 10/100BASE-T to 30 percent T1, excluding fiber cost).
Ethernet in the First Mile (EFM) Initiative
Metro solutions for Consumer/SOHO markets are different in the number of access sites involved,
and the price points that must be met. The requirements include very low-cost Internet access, highbandwidth video services as well as telephony. The key to getting the cost down and the
functionality up can be found in EFM (Ethernet in the First Mile) technology. EFM is an effort
within the IEEE to standardize a series of access technologies to cost effectively target the last mile.
The EFM study group was formed within the IEEE 802.3 (CSMA/CD) Working Group in
November 2000. Sixty-nine companies, including 3Com, Alloptic, Aura Networks, CDT/Mohawk,
Cisco Systems, DomiNet Systems, Intel, MCI WorldCom, and World Wide Packets, are
participating in the group. These Ethernet based access technologies can make use of both fiber and
twisted pair distribution plants. EFM technologies can take advantage of various Metro IP and
Metro Ethernet systems for transport back to the POP. EPON vendors are actively engaged in this
effort.
In addition to the IEEE study group, EPON vendors participating in other standards efforts
conducted within organizations, such as the Internet Engineering Task Force (IETF), ITU–
Telecommunications Standardization Sector (ITU–T), and the Standards Committee T1. There is
also a liaison with FSAN on this effort. The FSAN document (G.983) does not preclude non–ATM
protocols, and the FSAN document is broad in scope (covering many last mile issues). Much of
G.983 remains valid, and it could be that the IEEE 802.3 EFM group will focus on developing the
MAC protocols for EPON, referencing FSAN for everything else. This is the quickest path to an
EPON standard, and several big names, including Cisco Systems and Nortel Networks, are backing
EPON over APON.
Comparison of PON with xDSL
This section will compare PON systems with xDSL technologies and describe the issues associated
with each.
ATM is an ultrahigh-speed, one-size-fits-all, cell-based data transmission protocol that may be run
over many physical-layer technologies such as xDSL modems. These are attached to twisted-pair
copper wiring and transmit data at speeds of 1.5 Mbps to 9 Mbps downstream to the subscriber and
64 Kbps to 1.5 Mbps upstream, depending on the condition and distance of the copper line.
Telework and Collaboration with Broadband Access
Page 32 of 36
Asymmetric digital subscriber line (ADSL), for instance, offers users an always-on service, but its
maximum downstream and upstream speeds are ultimately limited by distance and the aging copper
infrastructure; typically, only speeds of 1.5 Mbps over 12 kft are achieved. If the customer is not
directly connected to a CO–based digital subscriber line access multiplexer (DSLAM), then an
expensive upgrade to an existing outside-plant DLC system is usually the only solution.
Very-high-speed DSL (VDSL) extends ADSL downstream speed to a potential 52 Mbps, with a
proportionately lower upstream speed, but offers a shorter distance range (1 kft to 3 kft) than ADSL.
However, this too requires expensive outside plant electronics installed in a cabinet that must survive
severe temperature variations.
In addition to the distance problem, xDSL technology has inherent interference problems, a liability
with copper-based technology. PONs cannot be interfered with by AM band radio and other radio
frequency interference (RFI)/electromagnetic interference (EMI) sources. XDSL is largely
considered to be a short-term broadband solution; since it can be easily installed without an
expensive outside-plant infrastructure build, the existing copper plant can be used. The PON system,
however, is believed to offer an ultimate, end-to-end broadband platform that is future-proofed.
BANDWIDTH AND DISTANCE CONSIDERATIONS
Network S er vi ce Distance Limit for Bandwidth Ranges
PO N
Asymmetric VDSL
Max. of 9-12 miles for 155 Mbps symmetrical
Max. of 1,000 ft. for up to 51.84 Mbps for downstream
Max. of 3,000 ft. for 12.96 to 25.92 Mbps for downstream
Max. of 4,500 ft. for 6.48 to 12.96 Mbps for downstream
Max. of 1,000 ft. for 3.24 to 6.48 Mbps for upstream
Max of 3,000 ft. for 1.62 to 3.24 Mbps for upstream
Max. of 4,500 ft. for 1.62 to 3.24 for upstream
Symmetric VDSL
Max. of 1,000 ft. for 19.44 to 25.92 Mbps
Max. of 3,000 ft. for 6.48 to 12.96 Mbps
Other Options (not necessarily PON) for Access
Optical Ethernet is a fourth-generation Layer 2 MAN/WAN technology. Optical Ethernet is a
connectionless packet technology. It runs Ethernet on dark fiber, SONET and optical rings. Optical
Ethernet can be configured on a point-to-point basis (emulating a circuit), or on a point-to-multipoint
basis (emulating frame relay star networks) or on a many-to-many basis (emulating a broadcast LAN
across a configured set of customer sites). These configurations can be used to interconnect Layer 3
routers and routing switches, and Ethernet switches.
Telework and Collaboration with Broadband Access
Page 33 of 36
Optical Ethernet is transparent to Layer 3 transport and routing protocols (including DNS (Domain
Name System) and DHCP (Dynamic Host Configuration Protocol) and related tools, so it can work
with legacy protocols, SNA and IPX. It is scalable from 10 Mbps to 10Gbps+ Ethernet. Three major
technology developments are key to optical Ethernet, and each can be used between enterprise sites,
to access the carrier POP or as a backbone technology, in either private or service provider networks.
The simplest form of optical Ethernet is to run Ethernet over dark fiber, and to use Ethernet switches
to connect locations. Various fiber-based solutions for single-mode and multimode dedicated fiber
have been defined by the IEEE 802.3; for example, SX up to 550 meters, LX up to 5 Km and ZX up
to 70 Km. Ethernet's next plateau will be 10 Gbps. A key difference from previous versions of
Ethernet is that l0-GigE supports two physical interfaces, one optimized for the LAN environment,
the other for the WAN. The WAN option allows 10-Gigabit Ethernet to be transparently transported
across existing OC-192 SONET infrastructures.
Dense Wave Division Multiplexing (DWDM: This networking technique which has been deployed
in carrier networks for a number of years. enables multiple wavelengths (each with 2.5 Gbps or more
capacity) to be concurrently supported on a single fiber pair configured in a point-to-point or ring
topology. Increasingly, DWDM networks are being deployed to provide wavelength services to
support applications such as channel extension and storage networks for mainframe environments,
ATM and video. While, to date, DWDM has been delivered through optical multiservice platforms,
DWDM interfaces are becoming available on routers. Enterprises will be able to use optical architectures that will support both DWDM platforms and routers on a single ring. In the longer term. 10Gbps Ethernet support and end-to-end wavelength services across the WAN will make DWDM even
more important.
Resilient Packet Ring (RPR): This technology enables distributed Ethernet switching across optical
rings running on fiber, lambdas or SONET pipes. RPR can be implemented in routers, Ethernet
switches and optical platforms. The "users"-individual PCs or servers, or aggregation devices such
as Layer 2 switches, Layer 3 routing switches or conventional routers - connect to RPR rings over a
standard Ethernet interface operating at 10/100/1000 Mbps. These interfaces support IEEE802.1q/p,
and users can be locally attached. if the RPR loops through the customer building, or remotely
attached using Ethernet over dark fiber.
These Metro IP systems can distribute IP intelligence out to the edge of the network using Resilient
Packet Ring technology. RPR is a resilient packet ring technology that combines the intelligence of
IP routing with the optical rings. Designed primarily for metropolitan area networks, this approach
delivers scalable Internet service, reliable IP-aware optical transport, and simplified network
operations. A packet ring network consists of two counter-rotating fiber rings. Each optical ring can
be used at the same time to pass both data and control packets doubling the effective bandwidth. The
ring is self-healing and has a recovery time of less than 50ms. (In contrast, SONET rings do not use
the second ring in normal traffic situatons, only for backup).
Cisco Dynamic Transport Technology (DPT) is based on the Spatial Reuse Protocol (SRP), a Ciscodeveloped MAC-layer protocol for ring-based packet internetworking. Cisco has submitted Spatial
Reuse Protocol (SRP), which is an open, freely available specification (RFC 2892), to the IEEE
802.17 Resilient Packet Ring (RPR) Working Group for consideration as the basis for the industry
Telework and Collaboration with Broadband Access
Page 34 of 36
standard. Cisco’s DPT uses Intelligent Protection Switching (IPS) L2 Path Restoration and enhanced
IPS and scalability to deliver high performance and efficiency.
SRP is used in Cisco's Dynamic Packet Transport (DPT) product line, which has been shipping for
more than two years, largely to cable modem service operators. Cisco's initial applications were for
ISP intra-POP rings. DPT gave service providers a way to keep OC-3 port counts low and helped
minimize the number of IP subnets within a site. DPT has helped lower interPOP connections and is
used by cable modem operators to connect head-ends with data centers. Cisco claimed in excess of
12,000 installed DPT ports at more than 160 customers With Cisco's SRP, transit packets take precedence over entering packets, so there's no packet loss on the ring itself, which is an advantage when
compared with Ethernet switches. Ordinary Ethernet switches lack a cut-through path for transit
traffic, exhibit varying packet loss throughout the network as traffic congests at each node.
Companies like Appian Communications, Cisco, Dynarc and Lantern Communications are
developing resilient packet ring (RPR) technologies, which replace Ethernet's media access control
(MAC) layer with a new ring-based MAC. Atrica and a few others are addressing the challenge with
ringbased solutions that preserve the Ethernet MAC. All these approaches promise service providers
the best of both worlds: Ethernet's low cost and packet data efficiency, plus SONET's ring structure
reliability and rapid restoral. The IEEE formed the 802.17 working group in 2000 to standardize
RPR. The IETF also has a working group called IP over RPR (IPoRPR).
Service providers like their existing SONET rings and for good reasons. Besides SONET's 50millisecond restoral, the rings reach more customers with less fiber than other topologies and require
fewer switch ports at busy hub sites. But SONET, which was designed for voice circuits, wastes
bandwidth when carrying packet traffic, and the spanning tree algorithm that guides most Ethernet
switches purposely breaks rings in order to prevent bridging loops. Packet over SONET (POS) offers
a partial solution, but only for point-to-point links. A better solution, according to the RPR
proponents, would be a new packet MAC that uses rings efficiently but also exhibits the resilience
and QOS of SONET.
Telework and Collaboration with Broadband Access
Page 35 of 36
In dual-ring topologies, SONET uses only one ring to carry live traffic; the other is reserved as a
backup. To increase fiber utilization, RPR will send traffic over both rings (in opposite directions)
during normal operation. An RPR ring will continue operating despite the loss of any node. In the
event of a fiber break or node failure, RPR will restore service at least as fast as SONET (50ms).
This will likely entail linking the two rings into one, so some traffic may get bumped. In some older
packet ring architectures, the source node removes unicast packets after they come all the way
around. With RPR, destination nodes remove their unicast packets, freeing downstream bandwidth
for reuse by other flows. Together with packet multiplexing and counter-rotating rings, destination
stripping more than doubles RPR's total throughput compared to SONET. RPR supports multicast
traffic: multicast packets travel once around the ring to reach every node. In contrast, mesh networks
must replicate multicast packets in order to reach all destinations. RPR will be fast enough to carry
Gigabit and 10-Gigabit Ethernet, but will also support lower data rates. New nodes may join the ring
without manual configuration. RPR will deliver QoS. RPR will allocate bandwidth to competing
flows on demand. RPR's primary mission is to make optical rings more efficient for packet traffic,
but the standardization effort has attracted some developers who want to see it do more, especially in
the realm of QOS and traffic control.
To provide end-to-end QoS SLAs, service providers can implement IP QoS features such as
classification, rate limiting, traffic shaping and traffic policing at the edge, and features like WRED
and MDRR/WFQ in the core. For Content Delivery Networks, the network can use integrated
content awareness for load balancing across CDN content engines (CEs) and for prioritizing traffic
based on URLs and cookies.
RPR can be viewed as a SONET ADM replacement or simply a high-performance packet MAC. The
rationale of the SONET replacement camp is the need to handle voice efficiently. In contrast, the
keep-it simple camp argues that TDM support will be RPR's Undoing by making it too complex.
Appian offers a hybrid approach. Its Optical Service Activation Platforms (OSAPs) dedicate some
channels on a SONET ring to packet traffic and leave others for ordinary TDM. This allows OSAPs
to share fiber rings with conventional SONET ADMs. Dynarc's implementation is more a blend than
a hybrid (Sweden). Channels, in turn, are mapped into TDM-like timeslots. Channel capacity can be
fixed, perhaps to tunnel a T1 through the network, or flexible to accommodate less sensitive flows.
Atrica rejected the RPR MAC in favor of a 10-Gbps Ethernet MAC over WDM wavelengths.
Furthermore, Atrica's Optical Ethernet Switches carry low speed TDM traffic in Ethernet frames. At
speeds from T1 (1.5 Mbps) to OC-12 (622 Mbps), Atrica uses Ethernet circuit emulation. Others are
developing chips for EPON and EFM aimed at noth TDM and Ethernet services.
Access Services
To provide high Bandwitdth Virtual Private Network services across Metropolitan Areas, service
providers are considering Ethernet on fiber, DWDM and RPR, or a combination of the three. The
demarcation between the user's environment and the network can be provided via an Ethernet User
Network Interface (UNI ). In the simplest case, customers directly access the POP over the
appropriate Ethernet on fiber UNI (usually GigE on single-mode fiber). The service provider may
provide a demarcation device at a customer premise or a multi-tenant unit, which provides multiple
10/100/I000-Mbps Ethernet UN Is and Gigabit Ethernet connectivity into the network. From there,
the traffic is carried over point-to-point fiber links, DWDM or RPR optical rings. These demarcation
devices allow multiple customers to share access bandwidth.
Telework and Collaboration with Broadband Access
Page 36 of 36