1. Tony Doyle a.doyle@physics.gla.ac.uk “GridPP2 Proposal”, GridPP7 Collab. Meeting, Oxford, 1 July 2003

2. Tony Doyle - University of Glasgow GridPP2 Proposal http://www.gridpp.ac.uk/docs/gridpp2/ ~30 page proposal + figures/tables + 11 planning documents: 15.Tier-0 16.Tier-1 17.Tier-2 18.The Network Sector 19.Middleware 20.Applications 21.Hardware Requirements 22.Management 23.Travel 24.Dissemination 25.From Testbed to Production A lot of work by everyone in a relatively short time…

3. Tony Doyle - University of Glasgow GridPP Mid-Term Review 1.What has been achieved in the first 18 months? Oversight Committee (May 2003): Conclusion "Excellent progress, management, control, reporting and thinking (both tactical and strategic), we look forward to an equally successful second half of the project as the investments made translate into middleware roll-out and benefits to the HEP experiments. It is difficult to imagine more impressive progress all round." 2.Development in the medium-long term? GridPP1 (09/01-08/04)Prototypeshort “Web to Grid” GridPP2 (09/04-08/07)Productionmedium “Prototype to Production” (09/07-) Exploitationlong-term “LHC exploitation”

4. Tony Doyle - University of Glasgow What impressed the Oversight Committee? (I) Identifiable progress via tasks and metrics “At the midpoint of the project, over half of the 182 tasks are completed and all of the 44 metrics are within specifications. The Project Map incorporates a complete Risk Register with 76 identified risks, reviewed on a regular basis.”

5. Tony Doyle - University of Glasgow Middleware Development Why were the Oversight Committee convinced? (II) What impressed the Oversight Committee? (II)

6. Tony Doyle - University of Glasgow Why were the Oversight Committee convinced? (III) Identifiable progress via testbed monitoring “All testbed sites can be said to be truly on a Grid by virtue of their registration in a comprehensive resource information publication scheme, their accessibility via a set of globally enabled resource brokers, and the use of one of the first scalable mechanisms to support distributed virtual communities (VOMS).” UK-wide development using EU-DataGrid tools (v1.47). Not yet robust, but sufficient for prototype development. See http://www.gridpp.ac.uk/map/ What were the Oversight Committee looking forward to? (I)

7. Tony Doyle - University of Glasgow Application Interfaces What were the Oversight Committee looking forward to? “benefits to the HEP experiments” What were the Oversight Committee looking forward to? (II)

8. Tony Doyle - University of Glasgow Where will we be in September 2004? Applications will be using the Grid for their Monte Carlo production. Typically low level usage (~5% of the CPU and ~1% of the storage). EDG will have completed its task of providing prototype middleware. EGEE will be in progress with more emphasis on a common infrastructure for a broader range of applications. Testing of LCG Phase 1 will be well underway. A prototype Tier-0 Centre will be linked to the Regional Tier centres based on the LCG releases. GridPP1 will be completing its task of developing and deploying a prototype Grid in the UK. A prototype Tier-1 Centre will exist based at RAL. Tier-2 centres will be being built up from hardware provided by the Institutes. The UK testbed will be fully integrated with the LCG testbed. The prototype Grid will have been rolled out to Institutes across the UK. Experiment applications will have been developed to various levels of functionality. Grid computing will be expected to provide for the requirements of all experiments.

9. Tony Doyle - University of Glasgow What will we need to do in September 2004? Full functionality production Tier-1 Centre needs to be developed. Tier-2 centres will need to be developed and integrated. The software development (verification, validation and testing) cycle will need to conform to industrial standards. EGEE will have assessed middleware. Dependent upon LCG for its operations centre. In the UK, the Grid Support Centre will have expanded its role. Need to incorporate anticipated wider participation, a move from Monte Carlo production to a full Physics Analysis environment The applications must be fully developed, tested and exercised in a production Grid environment. Scale? The total CPU capacity corresponds to ~7000 kSI2000 (or 20,000 PIII 1GHz processors) and 750 TBytes of disk storage in Sept. 2004. LCG will need to build the full facility needed by the LHC experiments including Tier-0 and Tier-1 resources at CERN. We will need to ensure that our deployment plans are fully interoperable especially where resources are shared with other disciplines.

10. Tony Doyle - University of Glasgow What do we plan to do? Build a Production Grid Components

11. Tony Doyle - University of Glasgow GridPP2 Timeline

12. Tony Doyle - University of Glasgow High Level Deliverables At the start of each year the Project Management Board will review these in order to specify that year’s target in terms that can be easily defined and monitored across the Grid.

13. Tony Doyle - University of Glasgow Tagged release selected for certification Certified release selected for deployment Tagged package Problem reports add unit tested code to repository Run nightly build & auto. tests Grid certification Fix problems Application Certification Build System Certification Testbed ~40CPU Production Testbed ~1000CPU Certified public release for use by apps. 24x7 Build system Test Group WPs Unit Test Build Certification Production Users Development Testbed ~15CPU Individual WP tests Integration Team Integration Overall release tests Releases candidate Tagged Releases Releases candidate Certified Releases Apps. Representatives From Testbed to Production

14. Tony Doyle - University of Glasgow GridPP2 Project Map Need to build this in: to identify progress…

15. Tony Doyle - University of Glasgow Projected Hardware Resources Total Resources: 20042007 (note x2 scale change)

16. Tony Doyle - University of Glasgow Experiment Requirements: UK only Total Requirement:

17. Tony Doyle - University of Glasgow CPU Estimates - “comparison” across (x,t) Estimated CPU Capacity required at CERN 0 100 200 300 400 500 1998199920002001200220032004200520062007200820092010 Jan 2000: 3.5K SI95 Other experiments LHC K SI2000 Estimated CPU Capacity required at CERN 0 1,000 2,000 3,000 4,000 5,000 1998199920002001200220032004200520062007200820092010 Jan 2000: 3.5K SI95 Other experiments LHC K SI95 Moore’s law – some measure of the capacity technology advances provide for a constant number of processors or investment CPU comparison considerations – 1.“historical precedent of underestimation?”: factor 2 increase in CERN estimates (over 3 years). 2.“meeting requirements of ALL experiments in UK?”: factor 4 to 2 3.“RAW-AOD-ESD-TAG”- “analysis” : previously largely un-estimated

18. Tony Doyle - University of Glasgow Tier Centres - Functionality RAW ESD AOD TAG “Interesting” Events List RAW ESD AOD TAG RAW ESD AOD TAG Tier-0(International) Tier-1(National) Tier-2(Regional) Tier-3(Local) Data Files Data Files Data Files TAG Data Files Data Files Data Files RAW Data File Data Files Data Files ESD Data Files Data Files AOD Data Event 1 Event 2 Event n

19. Tony Doyle - University of Glasgow Tier-1/A Services [FTE] Management [1.5] DataCPUUKMiddlewareExperiment DiskTapeFile System DeploymentSupport [1.5][2.5][0.5][2.0] Core Services [2.5]Operations [3.0] Security [1.0]Networking [0.5] High quality data services foremost requirement of a Tier-1 Centre A National and International Role UK Focus for International Grid development Established track record at RAL

20. Tony Doyle - University of Glasgow Tier-2 Centres [FTE] Four Regional Tier-2 Centres London: Brunel, Imperial College, QMUL, RHUL, UCL. NorthGrid: CLRC Daresbury, Lancaster, Liverpool, Manchester, Sheffield. ScotGrid: Durham, Edinburgh, Glasgow. SouthGrid: Birmingham, Bristol, Cambridge, Oxford, RAL PPD. Hardware resources are provided by Institutes GridPP provides manpower

21. Tony Doyle - University of Glasgow Meeting the Experiments’ Hardware Requirements Significant… Production Grid inc. Tier-2 resources needed…

22. Tony Doyle - University of Glasgow Tier Centres - Service Evolution

23. Tony Doyle - University of Glasgow Tier Centres - Manpower Requirements GridPP1GridPP2 FY 03-04 FY 05-06 GridPP1GridPP2 FY 06-07 FY 08-09 Exploit n GridPP2Exploit n FY 04-05 FY 07-08

24. Tony Doyle - University of Glasgow Middleware, Security and Network Service Evolution

25. Tony Doyle - University of Glasgow Middleware, Security and Network Service Evolution Information Services: strategic role within EGEE

26. Tony Doyle - University of Glasgow Application Interfaces - Service Evolution Currently 18 FTE working in experiment interface areas Encourage cross-fertilisation between collaborations Extend Grid across PP Proposed new activities 21 FTE GridPP2 request

27. Tony Doyle - University of Glasgow “Management” Conclusion: retain existing structure less “development” more “deployment” and “use”

28. Tony Doyle - University of Glasgow Summary Tables

29. Tony Doyle - University of Glasgow Roadmap: GridPP2 Proposal GridPP1GridPP2Exploitation

30. Tony Doyle - University of Glasgow Components: GridPP2 Proposal

31. Tony Doyle - University of Glasgow PPPARC Review Timeline: Projects Peer Review Panel (14-15/7/03) Grid Steering Committee (28-29/7/03) Science Committee (October 03) Components: GridPP2 Proposal (£23.1m) VerticalIntegrationVerticalIntegration

32. Tony Doyle - University of Glasgow c.f. GridPP6 Collab. Meeting: Proto-Proposal

33. Tony Doyle - University of GlasgowConclusion We request £23.1m to fund our three-year project, GridPP2 This will provide a Production Grid incorporating 1.access to the Tier-0 Centre at CERN and the LCG deployment releases 2.the UK Tier-1 Centre 3.integration of four distributed Tier-2 centres 4.technical development in Middleware, Security and Networking 5.Grid integration of the experiments The project is in direct support of PPARC's highest priority science programme, the LHC