The UK e-Science Core Programme Annual Report Year 1: April 2001 – April 2002 Contents Executive Summary 1 Introduction 2 The Core e-Science Programme Programme Team Structure of the Programme UK e-Science Grid Centre Projects Promotion of Middleware IRC Grid Projects Support for e-Science Projects International Collaboration and Outreach e-Science Network Requirements 2 2 2 3 4 5 6 7 8 9 Committees and Bodies within the Core Programme 10 Meetings Attended 11 Programme and Project Management 11 Finances 12 Conclusions 12 Executive Summary This report gives an overview of the first twelve months of activity of the Core e-Science Programme. A more detailed report of the activities is also available and is referred to occasionally in this document. The EPSRC, on behalf of all the Research Councils, and the Department of Trade and Industry are managing the programme. The Core Programme has been structured around six key elements: 1. Implementation of a National e-Science Grid based on a network of e-Science Centres 2. Promotion of Generic Grid Middleware Development 3. Interdisciplinary Research Collaboration (IRC) Grid Projects 4. Establishment of a support structure for e-Science Pilot Projects 5. Support for involvement in International activities and outreach activities 6. Support for e-Science networking requirements Each of these key elements is outlined and the status of the developments so far. A National Centre and eight Regional Centres have been created and a UK e-Science Grid has been initiated across these centres. It is essential that there is industrial collaboration and buy in, particularly on the development of the middleware to support e-Science activities. The Core Programme is proving successful in this regard, through industrial projects at the e-Science centres and other projects. During this first year of the programme there has been a shift in the industrial participation in Grid activities, with the many high tech companies such as IBM, Sun, HP, Oracle and others all having declared support for Open Source Grid middleware. The programme is also funding projects in middleware through an Open Call. Projects based on the four existing EPSRC computational Interdisciplinary Research Collaborations (IRC) have also been funded to enable them to consider the implications of Grid technology for their research directions. In addition, two collaborative projects combining the software technologies of the Equator and AKT IRCs with the MIAS application project have been funded. In effect these projects, each constitute a sort of ‘Grand Challenge’ pilot project in ‘e-Healthcare’. A Grid Support Centre has been created to support the e-Science pilot projects and other e-Science activities. The Centre provides telephone and email support for the use of grid middleware and related issues. It is also responsible for the digital certificate authorisation for e-Science applications. A Grid Network Team has been created to assist application developers in understanding network requirements. The team are also looking at the long-term network requirements for Grid applications. It is important to the success of the programme that it is sufficiently integrated with other related international activities. A number of funding opportunities to assist in this area have been put in place. The directorate have ensured that leading developers in this technology are involved in the programme either as technical advisors or through some similar route. The programme is seen as leading the way in Europe. Demonstrators have been developed which are now being used in educational and outreach activities. The demonstrators include applications from each of the Research Councils. The planned financial commitment level has been met for the first year. The programme is by nature complex and we provide an overview of the management mechanisms for the overall activity and the individual projects. Part of the co-ordination of projects involves the a database at the National e-Science Centre which is viewable by any one on the Internet and which also has an interface indicating what areas projects are involved in. 1 1. Introduction The e-Science Core Programme has OST funding of £15M. This sum is augmented by an allocation of £20M from the Department of Trade and Industry making a total of £35M. The programme is being managed by the EPSRC on behalf of all the Research Councils. As is common in DTI programmes, the DTI contribution of £20M requires a matching contribution from industry. It is also expected that there will be industrial contributions to the individual Research Council e-Science pilot projects making a total industrial commitment to the e-Science programme of well over £20M. The aim of the Core Programme is to support the e-Science pilot projects of the different Research Councils and work with industry in developing robust, ‘industrial strength’ generic grid middleware. Requirements and lessons learnt in the different e-Science applications will inform the development of more stable and functional grid middleware that can assist the e-Science experiments and be of relevance to industry and commerce. 2. The Core e-Science Programme Goal The overall goal of the e-Science Core Programme is to identify the key generic middleware requirements arising from the e-Science pilot projects. In collaboration with scientists, computer scientists and industry, the programme has a mandate to develop a framework that will promote the emergence of robust, industrial strength Grid middleware that will not only underpin individual application areas but also be of relevance to industry and commerce. The middleware developed within this programme must be to Open Source or to Open Standards. 2.1 The e-Science Core Programme Team The Director of the programme is Professor Tony Hey who is seconded to the EPSRC to direct the programme on behalf of all the Research Councils. Dr Anne E Trefethen acts as Professor Hey’s deputy and Dr James Fleming is the Associate Programme Manager at the EPSRC. Krys Bartoszewska and Carol Becker provide the programme support at EPSRC. The DTI half of the team consists of Ray Browne, who is the e-Science team leader at the DTI and who co-ordinates all the DTI activity within the programme and his colleagues Lee Vousden and Jackie Whalley. 2.2 Structure of the Core Programme The Core Programme has been structured around six key elements: 1. Implementation of a National e-Science Grid Testbed based on a network of Regional eScience Centres 2. Promotion of Generic Grid Middleware Development 3. Interdisciplinary Research Collaboration (IRC) Grid Projects 4. Establishment of a support structure for e-Science Pilot Projects 5. Support for involvement in International activities and outreach activities 6. Support for e-Science networking requirements We briefly discuss each of these activities below for each we present the goals and deliverables of that key element and the status to date. 2 2.3 The UK e-Science Grid and the e-Science Centres Goal There is an urgent need for the UK to develop a nucleus of people who have experience in the detail of running a Grid. A primary role for each Centre is therefore to donate a specific amount of computing and storage resource for use in the construction of a national e-Science Grid. The donated resources range from time on supercomputers and commodity clusters to databases and other repositories. Gaining experience in the use of digital certificates for single sign-on and authentication across the Grid is a key purpose of this work and sorting out problems with the security, firewalls and policy issues of each of these sites is a difficult task. In order to have the activity succeed in developing technologies that will go beyond the end of the programme and be taken up by industry, it is important to have industry engaged in the programme. There are two components to this involvement, education and understanding of the something of the technologies, and also the development itself. In addition each Centre has a responsibility to establish an appropriate regional outreach programme. This could take various forms – from a seminar series to joint projects and/or visits to relevant organizations in the region. The National Centre is funded more substantially for such activities, which include running training courses and a national seminar series. Outputs/Deliverables: 1. 2. 3. 4. A working UK e-Science Grid An appropriate-scale Regional Outreach programme. A national Outreach programme with training courses and a national seminar series. Full commitment of the budget to relevant collaborative industrial projects producing useful generic grid middleware 5. Securing significant additional funding, obtained by the Centres 6. Establishing good working relationships with PPARC Tier 1 and 2 Centres 7. Expanding the scope of the National e-Science Grid beyond the initial 10 university sites. 8. Industry engagement in development of software. 9. Technology transfer academic to industry and industry to academic. 10. Setting up and operating an e-Science Institute. Status Nine e-Science Centres have been established at the locations shown on the map of the UK in Figure 1. Figure 1: UK e-Science Grid Belf 3 National e-Science Centre (www.nesc.ac.uk) Belfast Regional Centre (www.qub.ac.uk\escience) Cambridge Regional Centre (www.cms.cam.ac.uk/grid.html) Cardiff Regional Centre (www.wesc.ac.uk) London Regional Centre (www.lesc.ic.ac.uk) North East Regional Centre Newca (www.neresc.ac.uk) North West Regional Centre (www.esnw.ac.uk) Oxford Regional Centre (e-science.ox.ac.uk) Southampton Regional Centre (www.e-science.soton.ac.uk) Hinx A National e-Science Centre has been established in Edinburgh, managed jointly by Glasgow and Edinburgh Universities. Eight other Regional Centres have been established – in Belfast, Cardiff, Manchester, Newcastle, Oxford, Cambridge, London (Imperial College) and Southampton - giving coverage across the UK. Manchester currently operates the UK’s national Supercomputer service. Figure 1 also shows the Rutherford and Daresbury Laboratory sites of CCLRC. These national laboratories are key sites of the UK e-Science Grid. The Hinxton site near Cambridge is also shown in Figure 1: Hinxton hosts the European Bioinformatics Institute (EBI), the Sanger Centre and an MRC Institute. This constitutes one of the major centres of genomic data in the world. It is therefore important that this site is linked to the e-Science Grid with sufficient bandwidth to support a number of e-Science bioinformatics projects. The National Centre in Edinburgh has also been funded to establish an ‘e-Science Institute’. This Institute is already organizing a programme of multidisciplinary research seminars covering a wide range of topics, with scientists and experts from all over the world. Their brief is to make the Institute an internationally known Centre for stimulating intellectual debate on all aspects of eScience. The Centres also have a centre-specific DTI budget (£3M NeSC: £1M Regional – see above) to develop suitable Grid Middleware projects collaborating with industry. This is part of the effort to engage industry, to ensure that ‘UK Plc’ understands the technology and is in a position to exploit the Grid in business and commerce. In addition, AccessGrid nodes have been established in each Centre to aid collaboration both within and outside the UK. The AccessGrid system was developed at Argonne National Laboratory in the USA and makes use of MBONE and Multicast technologies to provide a more natural videoconferencing experience between multiple sites that allows direct integration of Grid simulations and visualisation. This system allows easy interaction between the Centres and will be used to experiment with innovative ways of working and teaching. At the present time all but one of the centres has an operative AccessGrid node. Implementing the AccessGrid took longer than anticipated due to problems with multicast across the academic network. The Centres are contributing a heterogeneous collection of resources, including supercomputer and cluster computing systems as well as diverse data storage systems. (A full list of these resources can be found in the full version of this annual report.) These resources will constitute the UK e-Science Grid and will be used both as a test platform for new Grid middleware and as a resource available for Centre industrial projects. An Engineering Task Force (ETF) has been tasked with the implementation of this Grid (see Committees and Bodies section of report). The members of the Task Force are from the e-Science centres and are led by Dr David Boyd, CCLRC as part of the Grid Support Centre’s team. The Cambridge e-Science Centre has also been chosen to act as the NERC e-Science Centre, providing specific, focussed support for the NERC community. Each Centre produces a rolling quarterly report that acts as a report on the previous quarter and a plan for the upcoming quarter. They are also required to have a six monthly Centre review at which the programme’s Technical Advisory Group and/or e-Science Steering Committee members can participate (see Programme management section for detail). The Websites of all of the Centres are following a template developed at NeSC, thereby creating a UK e-Science Grid ‘branding’. Centre Projects The criteria for a Centre industrial project is that it is addressing a generic grid middleware issue, that it will result in an open source/open standard solution and of course that there is industrial collaboration. The extent of the contribution from industry is somewhat under the control of the 4 centre directors. The portfolio of projects from each Centre must have a 50% contribution from industry but any given project might be slightly above or below this figure. The proposals for projects are reviewed by members of TAG and the projects must be based on the following description. An e-Science Grid project must involve collaborative sharing of distributed resources – computing, data and/or facilities – in a secure manner. The Globus toolkit provides digital certificates through the Grid Security Infrastructure (GSI) for single sign-on to resources and user authentication. The Grid Support Centre at CCLRC has set up a Certificate Authority (CA) to issue digital certificates for this purpose. The Globus toolkit currently offers support for GSI and remote computing services, file access and fast data transfer. An alternative approach for data-centric problems could be to use the GSI package with Web Services, looking towards the Grid Service implementation of Globus 3.0, incorporating the Open Grid Standards Architecture (OGSA), now being taken forward for standardisation. (N.B. This is likely to be available from December 2002). To date 11 centre projects have been funded, and a further 9 are under consideration (decisions expected imminently). It is anticipated that a further 10-12 projects will be funded by July of 2002. The projects range from middleware for telemedicine and engineering optimisation to generic tools for understanding workflow. 2.4 Promotion of Grid Middleware Development Goal This is an essential component of the Core Programme. The challenge is to obtain Grid middleware requirements from all the e-Science pilots and deliver to each of the different Research Council communities Grid middleware and tools that the users welcome and use. At the same time, a dialogue with the relevant parts of the IT industry must be established to engage their interest in developing industrial-strength versions of the Grid middleware to support and enable e-Business opportunities. Industry also needs to be convinced of the value of an Open Standards approach to Grid middleware. The engagement with industry will be both directly via the Office of the Core Programme Director and each of the e-Science Centres. The Open Call for Grid middleware projects is important both for its openness and for its flexibility framework for assembling relevant consortia and projects. A major challenge will be to coordinate the software outputs from all these efforts as far as is reasonable. Deliverables/Outputs: 1. Requirements analyses from each of the different research communities 2. Identification and implementation of generic middleware component/tool valued by each research community 3. Commitment by major parts of the IT industry for an Open Standards approach to Grid middleware 4. Delivery of significant components of Grid middleware that are internationally valued and known to have been developed in the UK 5. Coordination of individual Centre and Open Call industrial projects to a reasonable extent Status In order to have the UK programme begin on an equal footing we have provided projects with a common starting point for Grid middleware. The initial Grid middleware selected is the same as that used by NASA in their Information Power Grid (IPG) and includes Globus, SRB (Storage Resource Broker) from San Diego, and Condor from Wisconsin. The IPG may be termed an ‘IntraGrid’ - since the NASA laboratories connected are all part of one organisation. In the case of the UK e-Science Grid, the Grid connects different universities with different IT policies, firewalls and so on. This is a good test of the basic Globus infrastructure and the digital certificate based security system. 5 The Core Programme is in discussions with major IT companies such as IBM, Sun, Oracle and Microsoft, as well as with the Globus, Condor and SRB teams concerning the future development of Grid middleware. In this respect, it is encouraging that both IBM and Sun have given strong endorsements to working with the Globus team to take forward the production of improved and robust Grid middleware. The software that will emerge will offer considerably more functionality than the present Grid middleware and will also be produced to industrial quality. A major task for the Core Programme is the capture of requirements for the Grid infrastructure from each of the e-Science pilot projects. These include computational, data storage and networking requirements as well as the desired Grid middleware functionality. In order that the projects do not dissipate their energies by fruitless re-explorations of common ground, the Core Programme has commissioned a number of reports on the present state of Grid middleware. Reports on Globus, SRB/Databases, and .NET are currently available at the NeSC website (see http://umbriel.dcs.gla.ac.uk/NeSC/general/technical_papers/). A Grid Architecture Task Force (ATF) has been set up, led by Malcolm Atkinson, Director of the National e-Science Centre in Edinburgh, to look at overall architectural directions for the Grid (see http://umbriel.dcs.gla.ac.uk/NeSC/general/teams/atf.html). They are tasked with producing a ‘eScience Grid Road Map’ for Grid middleware development. The ATF is tasked with identifying some specific short-term goals (given the pilot requirements) as well as identifying longer-term research issues. In addition, a Grid DataBase Task Force (DBTF), led by Norman Paton from Manchester, has been set up to examine the question of Grid middleware interfaces to Relational DataBase Management Systems and the federation of different data sources (see http://www.cs.man.ac.uk/grid-db/). Their preliminary ideas have been discussed with the Globus team, DataGrid, SRB and with IBM, Oracle and Microsoft. The DBTF has both a short term remit to look at developing an interface with some minimal useful functionality as soon as possible - and a longer term remit – to look at research issues beyond flat files and relational data. To that end they have produced a requirements document created by speaking with all active projects. Initial ideas from the DBTF point towards the implementation of a database interface in terms of a ‘Grid Services’ model along the lines of Web Services. This leads to the idea of the Grid middleware designed as a ‘Service Oriented Architecture’ with Grid services consumed by higher level applications. Reports from both task forces have been submitted to the Global Grid Forum as white papers, which were discussed at the last Global Grid Forum. It is important that at least a subset of standard protocols that go beyond the present Globus model are agreed as quickly as possible in order to ensure that the Grid middleware development in the application projects can proceed effectively. We intend to collaborate with the Globus team and assist in taking forward the open source implementation of these standards. The ATF and DBTF have written a draft of the eScience Grid Architecture Roadmap and there will be a workshop on April 18th where application developers and other middleware architects will have the opportunity to critique the proposed architecture. Without a community effort we cannot ensure that the architecture meets the requirements of the varied e-Science activities underway. The standards being adopted and developed within the scientific grid community over recent years have until recently taken a different path than those developed within the World Wide Web Consortium and other bodies active in the web services area. This potential divergence of technologies has caused concern, as there is clearly a need for these two sets of protocols to interface. In the last six or so months academics and industry have put forward an Open Grid Services Architecture (OGSA) that essentially provides a merging of the two strands. This grid services architecture will provide a unifying framework upon which all networked systems can be integrated, from high-performance computers to the most modest of commercial databases. The eScience community in the UK are actively involved in the specification and development of this architecture. The Core Programme has funded a joint centre project that will provide database 6 access within this architecture. The proposal is between NeSC, Newcastle and Manchester eScience centres and includes IBM and Oracle as industrial partners. In addition to the funding of the e-Science Centres, £5M (plus significant industrial contribution) is available through an ‘Open Call’ with no deadlines. One project has already been funded under the Open Call with three others under consideration and an anticipated four other proposals are expected in the near future. The projects for the Open Call must again be addressing a generic middleware issue – stretching the Grid in one direction or another. The proposals are reviewed by members of an e-Science ‘college’. 2.5 Interdisciplinary Research Collaboration Grid Projects The EPSRC in the UK has funded three, six-year, computer science (CS) oriented, Interdisciplinary Research Collaborations (IRCs). These are major projects that fund key CS research groups from a number of universities to undertake long-term research in three important areas. The Equator project, led by Tom Rodden from Nottingham, is concerned with technological innovation in physical and digital life. The Advanced Knowledge Technologies project (AKT) led by Nigel Shadbolt from Southampton is concerned with the management of the knowledge life cycle. Lastly, the DIRC project, led by Cliff Jones from Newcastle and Ian Sommerville from Lancaster, is concerned with the dependability of computer-based systems. A fourth IRC, jointly funded by EPSRC and the MRC, is the MIAS project led by Mike Brady from Oxford. This application focussed IRC is concerned with translating data from medical images and signals into clinical information of use to the medical profession. The Core Programme has provided further funding to these IRCs. Goal These IRCs were selected after an open competitive bidding process and represent a unique pool of expertise in these three key software technologies and in the important multidisciplinary application area of medical informatics. We want to leverage this computer science research activity and extend it to apply to the Grid. This may have happened naturally but the extra funding will ensure that the developments will be within the time of the programme and therefore the benefits to the broader audience will appear that much faster than otherwise. Deliverables/Outputs 1. 2. 3. 4. Leverage of outputs of leading edge computer science research into relevant Grid activities. Grid stretch of these technologies. ‘Grand Challenge’ pilots for e-Health Creation of IRC e-Science ‘community’ Status The Core Programme is funding projects with each of these IRCs to enable them to consider the implications of Grid technology for their research directions. In addition, we are funding two collaborative projects combining the software technologies of the Equator and AKT IRCs with the MIAS application project. In effect these projects, each constitute a sort of ‘Grand Challenge’ pilot project in ‘e-Healthcare’. Projects have been funded with AKT, Equator and MIAS, and joint projects AKT /MIAS and Equator/MIAS have also been funded. The DIRC proposal is now with reviewers and will be discussed at a TAG meeting on May 2nd. The IRC projects were reviewed by TAG with expert clinicians. All IRC projects are due to start in April. 7 2.6 Support for e-Science Projects Goal The e-Science projects in the UK are being developed at the same time as the middleware to support them. This leads to an inevitable tension and to the fact that application developers are being asked to use tools that have not been polished and in some cases are still under development. The application developers need support to use the middleware software. There is also a need to have a central ‘Certificate Authority’ in the UK to provide secure authority certificates for the use of Grid resources. Outputs/Deliverables 1. Interface to middleware developers providing smooth transition between releases for the UK. 2. Creation of a Grid starter kit (continually updated) 3. Provide information via web relevant to Grid technologies. 4. Act as the UK Certificate Authority. 5. Provide 9-5 telephone/email query service. 6. Educate systems level administrators in use/installation of middleware. 7. Facilitate the Engineering Task Force in implementation of UK e-Science Grid. Status In order to provide support for e-Science application projects, a Grid Support Centre (GSC) has been established, jointly funded by the Core Programme, BBSRC and PPARC (www.gridsupport.ac.uk). The GSC hosts teams located at CCLRC (RAL) and with associates at CCLRC (Daresbury) and Manchester and Edinburgh Universities. The GSC is charged with supporting the implementation and operation of the UK e-Science Grid and supporting associated Research Council Pilots, Centres and Projects. The Centre developed the UK ‘Grid Starter Kit’ which initially consists of the Globus Toolkit, Condor and the Storage Resource Broker middleware. The service provides regular upgrades to these core software systems, on-line help and advice for its installation and commissioning, the management of current Grid security certification system, coordination of the Engineering Task Force (ETF) and the first point of liaison with several key International groups. The support team are available to answer questions and resolve problems for Grid application developers across the UK on a 9 to 5 basis. Further roles of the team are to help educate the systems support staff at the UK e-Science Centres and to advise the Core Programme Directorate on all matters affecting the efficient and effective operation of the UK Grid and its connection to International collaborators. 2.7 International Collaboration and Outreach 2.7.1 International Collaboration Goal It is important to ensure that the UK e-Science community is actively communicating and collaborating with the international community. It is therefore desirable to encourage the development of an informed UK community on Grid technologies and provide funding for them to play an active role in the development of internationally agreed Grid protocols at the Global Grid Forum. 8 Deliverables/Outputs 1. Involvement in the Global Grid community and Global Grid Forum (GGF) 2. Collaboration with key international middleware development teams 3. Understanding of other national initiatives and directions 4. Engagement with non UK industrial ventures Status The Core Programme has funded a ‘GridNet’ network project which has a substantial travel budget for attendance of UK experts at relevant standards bodies – the Global Grid Forum, the IETF and W3C, for example. The last Global Grid Forum in February, saw a significant attendance and engagement by UK participants. The UK programme is also concerned to create meaningful links to international efforts represented by projects such as the EU DataGrid and the US iVDGL projects. We are therefore funding Grid fellowships for young computer scientists from the UK to participate in these projects. The National e-Science Centre is also tasked with establishing working agreements with major international centres such as Argonne National Laboratory, San Diego Supercomputing Center and NCSA in the USA and is active as a NPACI International partner. We are also looking to establish other international links and joint programmes In order to strengthen ties with the US research programme, members of the UK e-Science community took part in and NSF N+N Meeting in August 2001. A follow up meeting is planned at NeSC. Clearly the ties with CERN and the particle physics community in the UK and internationally are very important. The international links with EU projects are also important for the FP6 programme which is about to come on line. We also believe that the ESA, NASA connections will continue to be significant. The UK e-Science community is in a good position to take advantage of anticipated EU funding for e-Science and Grids. 2.7.2 Demonstrator Projects and Outreach Goal There is a need to educate the broader community about e-Science. Even to computational scientists it is not always clear what the Grid can offer to their science. The idea of the demonstrators is a short-term project that can use the present technology to illustrate the potential of the Grid in different areas. Outputs/Deliverables 1. 2. 3. 4. 5. Eleven active demonstrators of Grid technology Leading scientists who are able to present the demonstrators to their own communities Literature based on the demonstrators for outreach within industry ‘Bleeding Edge’ tests of the existing Grid middleware Industrial Grid Outreach – technology transfer Status The Core Programme has funded eleven ‘Grid Demonstrator’ projects. We have tried to select demonstrators across a range of applications representing each of the Research Councils. Examples include a dynamic brain atlas, a medical imaging project using VR, a robotic telescope project, automated chemical data capture and climate prediction. All demonstrators have been reviewed and most are in a final form. They will be used for both actual demonstrations of the technology at such events as the launch of the National Centre and will 9 be individually documented in a glossy brochure. It is anticipated that we will hold a number of ‘industrial’ days, as part of a DTI Grid outreach programme, later in 2002 at which the demonstrator projects will also be employed. On April 25th, Gordon Brown will be opening the National e-Science Centre, where the demonstrators will play a prominent role. As part of the community building and communication bridge, the Core Programme is organising an ‘All Hands’ meeting in which all e-Science projects will be encouraged to participate and the pilot projects will required to participate. The meeting is scheduled of September 2-4 and will be held in Sheffield. 2.8 Support for e-Science Network requirements Goal The UK e-Science application projects will rely on the UK universities network SuperJANET4 for delivering the necessary bandwidth. The backbone bandwidth of SuperJANET4 is now 2.5 Gbps and there is funding in place to upgrade this to 10 Gbps by mid 2002. Most application developers will not have had to consider network requirements previously and this can be a complex task. This task is not made easier by the ‘balkanisation’ of the UK network into an SJ4 backbone run by UKERNA and Metropolitan Area Networks run by university consortia. The goal, then, is to help the application scientists and the e-Science Centres understand their network requirements. A longer term goal is to consider what developments are required in order to ensure the UK network can support Grid applications of the future. Deliverables/Outputs 1. 2. 3. 4. Identify bottlenecks and potential bottlenecks. Analyse e-Science centre and project requirements Consider QoS issues for the Grid Long term strategy for network development Status The Core Programme has established a Grid Network Team (GNT). The team comprises academics from different areas of networking and members from UKERNA. The team have scheduled visits to the e-Science centres and have also visited those pilot projects that are ‘up and running’. A £500K project with UKERNA and CISCO has been funded to look at traffic engineering issues and another project is considering the question of bandwidth scheduling with the EU DataGrid project. Both of these network R&D projects are jointly funded by the Core Programme and PPARC. 3. Committees and Bodies set up within the Core e-Science Programme 3.1 Technical Advisory Group (TAG) The Technical Advisory Group is made up of UK and international academics and industrial members, who are able to advise the Director on the directions of the Programme, and who act as reviewers for the e-Science Centres and their projects. 3.2 Grid Network Team (GNT) As noted above the GNT, chaired by Professor David Huchinson, was established to advise the directorate and its Technical Advisory Group (TAG) on matters related to computer networking. There is a short-term role which is to assist e-Science application developers and e-Science Centres understand network requirements and identify network bottlenecks. A longer-term role is to identify 10 the directions that the UK needs to take to ensure that the network is as required by the growing eScience demand. 3.3 Architecture Task Force (ATF) As noted above the Architectural Task Force , led by Professor Malcolm Atkinson, was established to develop an architectural model for the UK's contributions to grid middleware. 3.4 Database Task Force (DBTF) The Grid DataBase Task Force (DBTF), led by Dr Norman Paton from Manchester, has been set up to examine the question of Grid middleware interfaces to Relational DataBase Management Systems and the federation of different data sources. 3.5 Engineering Task Force (ETF) The Grid Engineering Task Force was formed to guide the construction, testing and demonstration of a prototype UK e-Science Grid. It contains members from each of the ten UK e-Science Centres. The Task Force operates through several Working Groups. 3.6 Security Task Force (STF) The Security Task Force has been formed to help understand and prioritise the security issues of the Grid. The focus of the team is to provide the Director with suggested ways forward to tackle these issues. 4. Meetings Attended During the first twelve months of the programme it has been essential that the team meet with key industrial and academic players and with other funding agencies both in the UK and internationally. The meetings are of a variety of types, making ties with industry, international meetings, meeting with other funding agencies etc. The team members have attended over 200 meetings to this end (this number includes conferences and brief meetings). A detailed list of meetings attended is available in the detailed version of this report. 5. Programme and Project Management This is a complex programme with many inter-related projects and bodies. Figure 5 below illustrates this complexity by highlighting the connections between one pilot project and other elements of the programme. This same complexity exists when considering any one body or project and its relationship to the whole community. Many of these elements are beyond the ‘control’ of the Core Programme and at most we can try to ensure communication and awareness. In order to manage the Core Programme we are using a number of tools, including Microsoft Project. The Project has been defined at quite a detailed level (see figures below). This provides a picture of the programme at any one given point in time. To provide a higher-level view of the programme we have created an Action Plan (in Excel) – provided in Appendix 2. This provides the key events and has proved useful in determining likely budget requirements and effort. Management of the key elements of the programme is completed in slightly differing ways due to the nature of the elements. 5.1 Centre Management The e-Science Centres are required to provide a quarterly report/plan in which they indicate possible projects that are being developed, hence communicating that to other centres. There is a secure area on the NeSC website where papers and reports regarding the Centres are shared. 11 The Centre projects are reviewed by members of the TAG, and occasionally by other outside experts. The centre management teams manage the portfolio of projects but they are also coordinated through a project database web page on the NeSC site (see below). The Directorate also take an active role in several of the Centre management committees. Centre director meetings are held every 6-8 weeks where issues and opportunities can be discussed. Each centre has a six monthly review, beginning six months after the first industrial project is awarded. The review includes presentations from each of the active industrial projects and where the core team will review the financial and other details. The first of these reviews will be held in May at the Cambridge Centre. The reviews include members of the Core Programme team together with members of the TAG and e-Science Steering Committee. 5.2 Project Co-ordination A project mapping scheme has been developed. All projects are mapped on to the table of middleware components. The mapping identifies those components of Grid development that are included in any given project. This project mapping has been implemented as a web page. All e-Science projects are being stored in the database, indicating those components of middleware that they are actively engaged in developing, the funding agency and links to other projects. There is a visual representation of the database on the web page which indicates those areas that have well developed code, research code or none. A beta implementation of this mapping is available on the NeSC website. We intend including international as well as UK projects in the database which should provide a good overview for the community as a whole. The intention is that as middleware developers are considering proposals they will look at the web page and identify those areas where there is little or no activity. The All Hands meeting in September should ensure that there is cross communication of project plans and progress. 5.3 Demonstrator Management The management phase of the demonstrators is almost complete although there will be a need to ensure that they are maintained and are usable throughout the programme. Each demonstrator was reviewed during development by one or more of the Core Programme Team to ensure that the directions taken were as planned and matched the programme goals. The Grid Outreach programme using the demonstrators is being managed by the DTI but is also in collaboration with the EPSRC marketing group to ensure a same ‘branding’ of material. 5.4 Grid Support Centre The Grid Support Centre is reviewed quarterly. The Centre manager, Dr David Boyd, provides a status report and statistics of the number of requests for assistance that have been received, the number of hits and downloads of software from the website and so on. He also provides a set of milestones for the Centre for the upcoming quarter. The resources and effort are considered at each stage to ensure that the level of support is maintained. Both the GSC and ETF produce milestones and updates for the Directorate and the rest of the community as an integral part of the activity. The centre had its first such review on February 20th and the next is scheduled in May. 5.5 Open Call and IRCs Each of the open call and IRC projects will be reviewed together in a six monthly meeting which will be held along the same lines as the centre reviews. The first review meeting for these projects will take place in September. There will be a steering committee comprising the IRC directors and Core Programme directorate to monitor progress of the projects. 12 5.6 Risk Managment The Core Programme keeps an active document of risks and actions taken, or to be taken, to minimize the risk. This document is maintained regularly as the programme evolves and areas of risk develop. The programme regular report to the EPSRC RAC. 6. Finances As noted above the Core Programme is funded by the DTI and the Research Councils, with EPSRC managing the programme on behalf of the research councils. There are some of the elements of the programme that are funded jointly and others that are only funded from one source or the other. In the first year EPSRC has committed £9,052,350 and DTI £9,403,000. This is in line with the expected level of commitment for the first year of the programme. 7. Conclusions The Programme has seen a busy and productive year. All key activities have progressed as planned. The only area to have caused concern and moved at a slower rate than planned was the AccessGrid. This was due in part to a mismatch in multicast protocols across the network. The e-Science Centres are all up and running. The demonstrator projects are mostly complete. The Centre projects are beginning to take shape. The proposals for the Open call are just beginning to appear, perhaps slightly behind where we might expect them to be, however this is likely to be due to the effort being spent in bringing in Centre projects. The programme is quite complex and needs careful management to ensure success. The team are using a variety of management tools and are involved in regular reviews and meeting of the relevant parties. The project mapping promises to provide a good tool for project co-ordination and project information dissemination. The Programme is an internationally recognised programme and has been described as the leading ‘Grid’ programme in Europe. By leveraging international efforts, particularly from the US, the activity has been able to sprint off the starting blocks. The next twelve months sees a great deal of cross project communication activity, the building of the first draft of the architecture from the architecture roadmap, the Grid outreach programme to industry and the continuing development of industrial projects through the centres. 13