Oxford University e-Science Centre Grid Developments the Oxford

Oxford University e-Science Centre Grid Developments, the Oxford e-Science Centre and ITSS Paul Jeffreys Director OUCS Director Oxford e-Science Centre http: //e-science. ox. ac. uk/ paul. jeffreys@oucs. ox. ac. uk 1 1

Talk Outline Oxford University e-Science Centre • Summary of developments over last year – (Re-)Introduction to e-Science and the Grid – SR 2000 e-Science allocation – Oe. SC • Funds • Achievements • Team and Management • Access Grid – Race through Oxford projects – SR 2002 “e-Science” priorities • What does it mean for IT SS and OU? – What will be needed? 2 2

What’s in a name? ! Oxford University e-Science Centre • ‘Oxford e-Science Centre’ vs ‘Oxford Regional Grid Centre’ • However: – IBM Press Release quote (Aug. 2001): • “The driving force behind the evolutionary Grid Project is the global scientific community” – e-Science funding! • But in Oxford, much more than e-Science – …and clear change of emphasis in SR 2002 3 3

e-Science Oxford University e-Science Centre • John Taylor, Director General of the Research Councils, OST – “e-Science means science increasingly done through distributed global collaborations enabled by the Internet, using very large data collections, terascale computing resources and high performance visualisation” – "e-Science will change the dynamic of the way science is undertaken" 4 4

What is the Grid? Oxford University e-Science Centre • “The Grid is a software infrastructure that enables flexible, secure, coordinated resource sharing among dynamic collections of individuals, institutions and resources. ” [The Grid, eds. Foster & Kesselman] • “The Grid is an emergent infrastructure capable of delivering dependable, pervasive and uniform access to a set of globally distributed, dynamic and heterogeneous resources. It brings challenges of scalability, interoperability, fault tolerance, resource management and security. ” [UK e-Science Director] 5 5

Digital Mammography Oxford University e-Science Centre • Example to fire the imagination (Brady et al. )! – – – 1. 5 million screenings, 20% films lost, 80% false positive rate 40% increase; radiographer rather than radiologist -> need to develop standardised mammograms in federated database Enables ‘findonelikeit’ with clinical history Principle generic 6 6

SR 2000 e-Science Oxford University e-Science Centre • http: //www. research-councils. ac. uk/escience/ – “Research Councils invest £ 98 M for e-Science” • http: //www. dti. gov. uk/ost/link/news/issue 10_3. htm Foresight. LINK, Issue No 10 Summer 2001 – “The new £ 80 m LINK programme for e-Science Grid Technologies will spearhead the UK contribution to developing the next generation internet, supporting scientific research and e-business development in partnership with UK firms” “If UK industry wants a place in the world of business [Grid] applications, it has to get involved now - at the academic research stage” 7 7

SR 2000 e-Science Allocation Oxford University e-Science Centre DG Research Councils E-Science Steering Committee Director’s Awareness and Co-ordination Role Academic Application Support Programme Research Councils (£ 74 m), DTI (£ 5 m) PPARC (£ 26 m) BBSRC (£ 8 m) MRC (£ 8 m) NERC (£ 7 m) £ 80 m ESRC (£ 3 m) EPSRC (£ 17 m) CLRC (£ 5 m) Grid TAG Director’s Management Role Generic Challenges EPSRC (£ 15 m), DTI (£ 15 m) Collaborative projects Industrial Collaboration (£ 40 m) 8 8

Indicative Core Funding Breakdown Oxford University e-Science Centre 1. 2. 3. 4. 5. 6. 7. 8. Grid Centres Grid Middleware Grid IRC Projects Grid Support International GNT Demonstrators Pilots £ 7. 0 M + £ 2. 5 M £ 3. 0 M £ 11. 5 M + £ 2. 5 M + £ 1. 0 M + £ 2. 0 M £ 1. 5 M £ 0. 5 M £ 1. 0 M ______ £ 20 M + DTI £ 1. 0 M £ 1. 5 M ______ £ 15 M + OST 9 £ 11. 0 M £ 2. 5 M £ 4. 0 M £ 1. 5 M £ 8. 0 M ______ £ 27 M Industry 9

UK Grid Network Oxford University e-Science Centre £ 470 k to ‘pass go’ £ 1 M pot – matched by industry Oxford University Investment: Associate Directors (RDF £ 85 k), plus 3 posts phased (RDF £ 218 k), plus some subsidised effort Edinburgh Glasgow DL Belfast Newcastle Manchester Oxford Cardiff RAL Cambridge London Hinxton Soton Oxford has received £ 8 M and rising! 10 10

Oe. SC ‘Objectives’ Oxford University e-Science Centre • Establish Oxford as regional centre on national Grid – Thereby establish Grid connections for our researchers – Make our resources available on the Grid (part of the deal!) – Aim. . to become ‘Centre of Excellence’ -- NB SR 2002 • Support groups throughout University undertaking national and international e-Science projects (and other Grid activities), collaborate with Oxford Brookes, and link with companies – Provide physical infrastructure – Provide support infrastructure: - registration, certificate authorisation, training, documentation, security, services – Share development, coordinate and optimise across projects – Disseminate • Commission ‘intranet Grid’ – Share resources across university 11 11

Status of Oxford e-Science Centre Oxford University e-Science Centre • Formally launched 17 August – Associate Directors: David Gavaghan, Mike Giles, Mark Sansom, Jim Davies – Web site … http: //e-science. ox. ac. uk – Administrative Centre and Point-of-Contact (OUCS) sue. crowley@oucs. ox. ac. uk – Strong connections with Business Liaison Unit (from the outset) – Connections established with OII and SBS – Technical Development: • Access Grid installed in Com. Lab - and shortly in hospital-area • ‘Gate-keeper’ in OUCS running GLOBUS • 8 cpu Linux cluster running Condor scheduling – Management Structure established – Building close relationship with RAL (A 34 Corridor) – OUCS team built within Research Technologies Section, in Information and Support Group – Oxford University is online! 12 12

Oe. SC Management Oxford University e-Science Centre • Substantial/important activity; needs appropriate structure – Management Board • Overall ‘responsibility’ for delivery and operation of functional e-Science Centre, resources invested, accountability, outside relations and integration, and overseeing all e-Science and Grid activities in OU – Technical/Development Board • High level management of e-Science/Grid activities, internal resources invested on Oe. SC, coordination – Technical Committee • Detailed internal technical management and organisation – User Committee • Forum for Users working on e-Science and Grid projects, and for dissemination across the University 13 13

Access Grid Oxford University e-Science Centre • Collection of resources that support formal and informal group-to-group interaction across the grid • Supports large-scale distributed meetings and collaborative work sessions 14 14

Access Grid Technology Oxford University e-Science Centre • Access Grid Collaboration – Enable collaborative work at dozens of sites worldwide, with strong sense of shared presence – Combination of commodity audio/video tech + Grid technologies for security, discovery, etc. – 40+ sites worldwide, number rising rapidly – Clear area of interest to OII – Second installation (Churchill? ) Presenter mic Presenter camera Ambient mic (tabletop) http: //www. accessgrid. org Audience camera 15 15

Oxford Grid Project Overview Oxford University e-Science Centre • Projects under different headings – Research Council Related • EPSRC • Data. Grid • Other RC and IRC – Oe. SC Industrial • Company has to commit itself to Grid • Purpose is to ‘effect the connection’ and to disseminate • Much more information available…! 16 16

EPSRC e-Science Projects Oxford University e-Science Centre • EPSRC – Grid Enabled Optimisation and Design Search for Engineering • Aim to develop a new way of optimising engineering design through the use of a grid enabled design search service, and focusing on CFD – Mike Giles – The Reality. Grid – a tool for investigating condensed matter and materials • Aim to use Grid technology to closely couple high performance computing and visualisation (steering) to create an environment for modelling to be compared and to integrate with experimental data – Adrian Sutton – Distributed Aircraft Maintenance • Aim to build a generic grid testbed for distributed diagnostics of aircraft engines on a global scale – Lionel Tarassenko 17 17

Data. Grid Projects Oxford University e-Science Centre • Two projects each involving the use of computer science to underpin the implementation of Grid applications software – Grid security - develop provably correct security protocols to ensure secrecy, integrity and authenticity - Collaborators: OUCL (Gavin Lowe, Bill Roscoe, David Gavaghan), RAL, EU Data. Grid Security Coordination group – Data Management – with functionality checked • develop a distributed handling and job-submission system for very large datasets arising in experimental particle physics and use software engineering techniques to ensure that the implementation has the correct functionality – Collaborators: OUCL (Andrew Martin, Jim Davies, David Gavaghan, Todd Huffman, Ian Mc. Arthur), RAL, EU Data. Grid WP 2 18 18

Other R. C. Projects(1) Oxford University e-Science Centre • • • PPARC – Grid. PP: National Centre and 4 regional Centres – Todd Huffman, Ian Mc. Arthur, Tony Weidberg and many others EPSRC-MRC Interdisciplinary Research Consortium – – – ‘From Medical Signals and Images to Clinical Information’ Joint EPSRC/MRC IRC in Medical Imaging and Signals Awarded £ 500 K to develop and maintain the necessary Grid infrastructure to support e-Science activities between the IRC consortium and the associated Regional e-Science Centres BBSRC – Mike Brady, Lionel Tarassenko, Alison Noble, Steve Smith 1. A Grid Database for Biomolecular Simulations (£ 700 k + 10%) – Simulations, creating a federated database, with data mining – Mark Sansom, Paul Jeffreys – David Stuart (WTCHG) representing Oxford Component 2. Molecular Consortium; Structural Biology 19 19

Other R. C. Projects(2) Oxford University e-Science Centre • MRC – Project looking at very large clinical trials of cancer utilising genomic information; marrying genetic informations and clinical trial data – Prof David Kerr (Clinical Pharmacology and Therapeutics) • Reached the final stage of the MRC call • NERC – Closely related to ‘Climate. Prediction’ project discussed below • Wellcome – Working to establish Wellcome Grid Centre! • …also other projects underway with e-Science component 20 20

Oe. SC Industrial Projects… Oxford University e-Science Centre • Collaborative Visualisation and Computational Steering • Project Leader: Ken Brodlie (Computer Science, Leeds University) • Industrial Partners: NAG, Streamline Computing • Successful bid to Open Call – A key generic enabling technology for Grid computing is collaborative and remote visualisation and computational steering – This project will address this in the context of steady and unsteady three-dimensional data arising from CFD calculations at Oxford and Imperial College, and the heart-modelling project described above – The results will also be relevant to any scientific computations on an underlying computational mesh, using finite volume or finite element 21 21

…Oe. SC Industrial Projects … Oxford University e-Science Centre • Videoworks • Project Leader: David Shotton, Zoology Department • Industrial Partners: IBM, Informix, Virage • Funded -> £ 327 k – Aim to create a unified suite of generic and scalable video e-services specifically for the handling and analysing scientific digital video data files located in distributed file stores over high bandwidth academic Grid networks, with a particular emphasis on the usefulness of these video e-services for biomedical research 22 22

… Oe. SC Industrial Projects… Oxford University e-Science Centre • Remote use of scientific Instrumentation – – • Project Leader: David Cockayne (Materials) • Industrial Partner: JEOL • Submitted and awaiting response! (£ 250 k) Micro-analytical instrumentation, such as transmission electron microscopes, electron probes and surface analysis equipment is becoming increasingly expensive, increasingly sophisticated Consolidation of infrastructure seems both inevitable and desirable Many instruments can now be operated (at least in part) remotely Purpose is to develop a capability for satisfying this need in electron optical instrumentation, and to put in place a demonstration national grid of access to sophisticated electron optical equipment 23 23

… Oe. SC Industrial Projects … Oxford University e-Science Centre • Climate Prediction • Project Leaders: Myles Allen, Dave Stainforth (Atmospheric Physics) • Industrial Partners: IBM, RSL, up to 2 million PC owners worldwide • Submitted last week £ 320 k – Recent studies have shown that atmosphere-ocean general circulation models (AOGCMs) are capable of simulating some largescale features of present-day climate and recent climate change with remarkable accuracy – Aim is to develop techniques that will allow the use of up to 2 million home and business PCs to provide the first fully probabilistic 50 year forecast of human induced climate change – New challenge of collecting (relatively) large data samples 24 24

…Oe. SC Industrial Projects Oxford University e-Science Centre • e. Diamond (Digital Mammography) • Will be submitted to DTI as ‘special’ on Friday • Mike Brady, David Gavaghan, Paul Jeffreys • Expect to be £ 3 M, but only need modest amount from Oe. SC • Heterogeneous Workload Management and Grid Integration • IBM researcher on secondment to Oxford • Basically receive money for two staff! • Submit to DTI on Friday – End to end response time monitoring and differentiated service applications running within Virtual Organisation – Applied to e-Diamond 25 25

Oe. SC Projects under development… Oxford University e-Science Centre • Computational Drug Discovery • Project Leader: Graham Richards, Mark Sansom • Industrial Partners: Inhibox, Wellcome Trust? – Following screensaver project, more computationally demanding calculations on binding of small molecules to proteins using the Grid • Whole-heart Modelling • Project Leader: Denis Noble, David Gavaghan • Industrial Partners: The Wellcome Trust, possibly others – Biological function arises from a complex interplay between processes at all levels of organisation -> comprehensive biological simulation of the whole-heart to be undertaken in a grid environment 26 26

…Oe. SC Projects Under Development Oxford University e-Science Centre • Fully Accessible Video for e-Learning • Oxford/Cambridge (Stuart Lee) • £ 412 k • “Test water”… – An e-Learning project to develop the use of high quality video-based resources for teaching, enabling random access to video data files while operating at streaming speeds – Enables accurate access to precise points within datasources (eg down to sentences) • Library Services • Matthew Dovey, David Price – Looking for overlapping areas of interest, eg access to heterogeneous databases, methods for locating protocol gateways and information services, and distributed library architectures 27 27

Is the Grid Niche or Mainstream? Oxford University e-Science Centre • ‘A major development in 2001 has been the endorsement of the concept of the Grid by Microsoft, Sun and IBM’ (SR 2002 bid) • Irving Wladawsky-Berger (Lead for IBM Corporate on Grid) – ‘Grid computing is a set of research management services that sit on top of the OS to link different systems together’ – ‘We will work with the Globus community to build this layer of software to help share resources’ – ‘All of our systems will be enabled to work with the grid, and all of our middleware will integrate with the software’ • Dave Turek (IBM’s Vice President of LINUX) – ‘IBM's Global Services division plans to build $4 billion worth of such [Grid] "computer farms" around the globe’ • Ne. SC opened by Gordon Brown • Tony Blair’s talk … 28 28

Blair’s speech on British Science Oxford University e-Science Centre http: //politics. guardian. co. uk/speeches/story/0, 11126, 721029, 00. html • “The emerging field of e-science should transform this kind of work. It's significant that the UK is the first country to develop a national e-science grid, which intends to make access to computing power, scientific data repositories and experimental facilities as easy as the web makes access to information. One of the pilot e-science projects is to develop a digital mammographic archive, together with an intelligent medical decision support system for breast cancer diagnosis and treatment. An individual hospital will not have supercomputing facilties, but through the grid it could buy the time it needs. So the surgeon in the operating room will be able to pull up a high-resolution mammogram to identify exactly where the tumour can be found. “ 29 29

SR 2002(1) Oxford University e-Science Centre • SR 2002 – distinct change of emphasis 1. e-Learning • Virtual Laboratories, Access. Grid and other Grid-enabled technologies 2. e-Government • Many of the heterogeneous data collections of Government Departments should be federated using Grid middleware 3. Medical and Healthcare Informatics • Grid middleware to maintains privacy and trust; need to strengthen links with NHS activities 4. e-Environment • Grid infrastructure needs to be applied to a wide variety of environmental problems ranging from monitoring endangered species or traffic flows to flood warning systems or pollution monitoring 30 30

SR 2002(2) Oxford University e-Science Centre 5. e-Business – By the end of the SR 2002 period, robust, industrial-strength Gridmiddleware will be available -> ‘Virtual Organisations’ -> multi-company business opportunities 6. International e-Science collaboration – e-Science …s UK programme fits well into the EU’s European Research Area initiative 7. Military applications of Grid technologies – Joint programme should be established with the MOD 8. Security – Vital area of concern for the Grid and requires a focused subprogramme of R&D looking at the needs of the different application areas 31 31

SR 2002 OST resources Oxford University e-Science Centre SR 2000 Programme SR 2002 15. 0 Core 20. 0 (20. 0) (DTI) (25. 0) 8. 0 BBSRC 15. 0 CLRC 9. 3 17. 0 EPSRC 19. 5 3. 0 ESRC 8. 5 8. 0 MRC 10. 0 7. 0 NERC 10. 0 26. 0 PPARC 45. 0 89. 0 Total 137. 3 32 32

SR 2002 conclusions Oxford University e-Science Centre • More money requested for e-Science than in SR 2000 – Up to now Oxford University received > £ 8 M and growing • Activity is now mainstream – We are building a special relationship with IBM • Different emphases from SR 2000 – Clearly advantage to Oxford – Healthcare prominent! • We need to be preparing for SR 2002 – Oe. SC – building projects in the key areas and competing! – Preparing to bid against the RC funds 33 33

Implications for ITSS Oxford University e-Science Centre • Role of Computing Services and IT Support in Infrastructure – Generally understood that the Grid will form part of the basic IT infrastructure of the modern University • Meeting May 1, Ne. SC – e-Science Developers/Computing Services – ‘Goal is to have the advantages of the infrastructure, but not visible to user’ • Oxford uniquely placed – and leading • Computing Services becoming more important as move from Level 1 Grid (basic Grid functionality for one-off tests) to Level 2 Grid (persistent, robust, 24 X 7) • Majority of e-Science money in Oxford invested in applications – How do we support the researchers, who generally are not Grid experts? 34 34

What will it mean for IT SS (1) ? Oxford University e-Science Centre • Our IT users – “I need a digital certificate to access my collaborators’ computing resources” • Authentication by University of Oxford – “I need authorisation to use the database and Beowulf cluster at. . ” • Authorisation through their collaboration, but coordinated and facilitated here – “I need to be able to use my digital certification to launch jobs into the ‘Astro. Grid” • Means variety of things: – Knowing how to send secure requests with digital certificates – How to launch jobs (different schedulers) • Substantial help from Oe. SC will be available – “I need to advertise the resources here in Oxford to my collaboration (or I need to find out what resources are available)” • Need to be able to advertise/interrogate MDS GIIS/GIS services 35 35

What will it mean for IT SS (2) ? Oxford University e-Science Centre • Collaborators from outside Oxford wishing to use facilities owned by our users – “A collaborator based at … needs to access the OU Super. Computer and database” • Your user will need a ‘Grid Client’ or gatekeeper (could be central one or one set up locally) – This will need to receive his/her collaborator’s digital certificate • He/she will need to be authorised to use the facilities • You will need to take care of firewall and other security issues – “My collaborator has launched a job which requires facilities at Oxford, but it is running slowly/does not work” • At least will need to know how to route questions! 36 36

Likely questions from IT SS (1) Oxford University e-Science Centre • "How much am I going to need to know? “ • • – Not need to know intricacies of the Grid (unless wish to!) – Basic understanding will probably be necessary – Some exciting opportunities! "Where can I get help? “ – Oe. SC point of information: oesc@oucs. ox. ac. uk • Routed through RT to expert • All learning together to some extent! • Applications will build up considerable expertise which must be able to draw upon “How much of an additional workload is the introduction of the Grid is likely to impose upon me? ” – Hard to be sure – At the moment – specifically related to scientific applications, but could be much broader – How would you answer the question if asked about WWW? – Certainly true that expertise being built up across most Universities – Perhaps will mean closer association with application teams 37 37

Likely questions from IT SS (2) Oxford University e-Science Centre • “Should I be responsible for supporting users’ collaborators? ” – Ideally it should all balance out! – I suspect for some time it will operate through well defined mini. Grids (eg the Dame Grid) – The project itself will offer support • “How can I be kept informed and/or contribute? ” – email jon. hillier@oucs. ox. ac. uk – Training will be made available – Partnership needed 38 38

Summary Oxford University e-Science Centre • e-Science and the Grid are now mainstream – Will have profound effect on the University • Oxford has extremely exciting Grid projects • Oe. SC is one of the main centres within a world leading national Grid installation • Opportunity for OU to really lead the way in terms of creating a persistent, robust IT infrastructure – Must be done in partnership • Collaboration between IT SS, Oe. SC, OUCS and application researchers will be needed 39 39

Conclusions Oxford University e-Science Centre • • SR 2000 investment in e-Science has been successful Oe. SC is well advanced, managerially and technically Oxford has been very successful in bidding against RC funds The Industrial Projects are well developed and we will be one of the first Centres to allocate our funds! Over the last year, the Grid has become mainstream Definite change of emphasis in SR 2002 e-Science bid Oxford should be preparing now. . There will be profound implications for the University and for IT SS! 40 40