Presentation is loading. Please wait.

Presentation is loading. Please wait.

Tony Doyle GridPP – Project Elements UK e-Science All Hands Conference, Sheffield 3 September 2002.

Similar presentations

Presentation on theme: "Tony Doyle GridPP – Project Elements UK e-Science All Hands Conference, Sheffield 3 September 2002."— Presentation transcript:

1 Tony Doyle GridPP – Project Elements UK e-Science All Hands Conference, Sheffield 3 September 2002

2 Tony Doyle - University of Glasgow GridPP – Project Elements From Web to Grid… e-Science = Middleware LHC Computing Challenge Infrastructure –Tiered Computer Centres –Network BaBar – a running experiment Non-technical issues …Building the Next IT Revolution UK GridPP EU DataGrid –Middleware Development –Operational Grid DataGrid Testbed Status: 25 Jun :38:47 GMT GridPP Testbed Grid Job Submission Things Missing, Apparently… …From Grid to Web

3 Tony Doyle - University of Glasgow GridPP EDG - UK Contributions Architecture Testbed-1 Network Monitoring Certificates & Security Storage Element R-GMA LCFG MDS deployment GridSite SlashGrid Spitfire… Applications (start-up phase) BaBar CDF/D0 (SAM) ATLAS/LHCb CMS (ALICE) UKQCD £17m 3-year project funded by PPARC CERN - LCG (start-up phase) funding for staff and hardware... £3.78m £5.67m £3.66m £1.99m £1.88m CERN DataGrid Tier - 1/A Applications Operations

4 Tony Doyle - University of Glasgow Provide architecture and middleware Use the Grid with simulated data Use the Grid with real data Future LHC Experiments Running US Experiments Build Tier-A/prototype Tier-1 and Tier-2 centres in the UK and join worldwide effort to develop middleware for the experiments GridPP

5 Tony Doyle - University of Glasgow Who are we? Nick White /O=Grid/O=UKHEP/ White member Roger Jones /O=Grid/O=UKHEP/ Jones member Sabah Salih /O=Grid/O=UKHEP/ Salih member Santanu Das /O=Grid/O=UKHEP/ Das member Tony Cass /O=Grid/O=CERN/ Cass member David Kelsey /O=Grid/O=UKHEP/ Kelsey member Henry Nebrensky /O=Grid/O=UKHEP/ Nebrensky member Paul Kyberd /O=Grid/O=UKHEP/ Kyberd member Peter Hobson /O=Grid/O=UKHEP/ R Hobson member Robin Middleton /O=Grid/O=UKHEP/ Middleton member Alexander Holt /O=Grid/O=UKHEP/ Holt member Alasdair Earl /O=Grid/O=UKHEP/ Earl member Akram Khan /O=Grid/O=UKHEP/ Khan member Stephen Burke /O=Grid/O=UKHEP/ Burke member Paul Millar /O=Grid/O=UKHEP/ Millar member Andy Parker /O=Grid/O=UKHEP/ member Neville Harnew /O=Grid/O=UKHEP/ Harnew member Pete Watkins /O=Grid/O=UKHEP/ Watkins member Owen Maroney /O=Grid/O=UKHEP/ Maroney member Alex Finch /O=Grid/O=UKHEP/ Finch member Antony Wilson /O=Grid/O=UKHEP/ Wilson member Tim Folkes /O=Grid/O=UKHEP/ Folkes member Stan Thompson /O=Grid/O=UKHEP/ Stan Thompson member Mark Hayes /O=Grid/O=UKHEP/ Hayes member Todd Huffman /O=Grid/O=UKHEP/ Todd Huffman member Glenn Patrick /O=Grid/O=UKHEP/ N Patrick member Pete Gronbech /O=Grid/O=UKHEP/ Gronbech member Nick Brook /O=Grid/O=UKHEP/ Brook member Marc Kelly /O=Grid/O=UKHEP/ Kelly member Dave Newbold /O=Grid/O=UKHEP/ Newbold member Kate Mackay /O=Grid/O=UKHEP/ Mackay member Girish Patel /O=Grid/O=UKHEP/ D. Patel member David Martin /O=Grid/O=UKHEP/ J. Martin member Peter Faulkner /O=Grid/O=UKHEP/ Faulkner member David Smith /O=Grid/O=UKHEP/ Smith member Steve Traylen /O=Grid/O=UKHEP/ Traylen member Ruth Dixon del Tufo /O=Grid/O=UKHEP/ Dixon del Tufo member Linda Cornwall /O=Grid/O=UKHEP/ Cornwall member /O=Grid/O=UKHEP/ Li member Paul D. Mealor /O=Grid/O=UKHEP/ D Mealor member /O=Grid/O=UKHEP/ A Crosby member David Waters /O=Grid/O=UKHEP/ Waters member Bob Cranfield /O=Grid/O=UKHEP/ Cranfield member Ben West /O=Grid/O=UKHEP/ West member Rod Walker /O=Grid/O=UKHEP/ Walker member /O=Grid/O=UKHEP/ Lewis member Dave Colling /O=Grid/O=UKHEP/ D J Colling member Alex Howard /O=Grid/O=UKHEP/ Howard member Roger Barlow /O=Grid/O=UKHEP/ Barlow member Joe Foster /O=Grid/O=UKHEP/ Foster member Alessandra Forti /O=Grid/O=UKHEP/ Forti member Peter Clarke /O=Grid/O=UKHEP/ Clarke member Andrew Sansum /O=Grid/O=UKHEP/ Sansum member John Gordon /O=Grid/O=UKHEP/ Gordon member Andrew McNab /O=Grid/O=UKHEP/ McNab member Richard Hughes-Jones /O=Grid/O=UKHEP/ Hughes-Jones member Gavin McCance /O=Grid/O=UKHEP/ McCance member Tony Doyle /O=Grid/O=UKHEP/ Doyle admin Alex Martin /O=Grid/O=UKHEP/ member Steve Lloyd /O=Grid/O=UKHEP/ admin John Gordon /O=Grid/O=UKHEP/ Gordon member/O=Grid/O=UKHEP/ White/O=Grid/O=UKHEP/ Jones/O=Grid/O=UKHEP/ Salih /O=Grid/O=UKHEP/ Das/O=Grid/O=CERN/ Cass /O=Grid/O=UKHEP/ Kelsey/O=Grid/O=UKHEP/ Nebrensky/O=Grid/O=UKHEP/ Kyberd/O=Grid/O=UKHEP/ R Hobson/O=Grid/O=UKHEP/ Middleton/O=Grid/O=UKHEP/ Holt/O=Grid/O=UKHEP/ Earl/O=Grid/O=UKHEP/ Khan/O=Grid/O=UKHEP/ Burke/O=Grid/O=UKHEP/ Millar/O=Grid/O=UKHEP/ /O=Grid/O=UKHEP/ Harnew/O=Grid/O=UKHEP/ Watkins/O=Grid/O=UKHEP/ Maroney/O=Grid/O=UKHEP/ Finch/O=Grid/O=UKHEP/ Wilson/O=Grid/O=UKHEP/ Folkes/O=Grid/O=UKHEP/ Stan Thompson /O=Grid/O=UKHEP/ Hayes/O=Grid/O=UKHEP/ Todd Huffman/O=Grid/O=UKHEP/ N Patrick/O=Grid/O=UKHEP/ Gronbech/O=Grid/O=UKHEP/ Brook/O=Grid/O=UKHEP/ Kelly/O=Grid/O=UKHEP/ Newbold /O=Grid/O=UKHEP/ Mackay/O=Grid/O=UKHEP/ D. Patel/O=Grid/O=UKHEP/ J. Martin/O=Grid/O=UKHEP/ Faulkner/O=Grid/O=UKHEP/ Smith /O=Grid/O=UKHEP/ Traylen/O=Grid/O=UKHEP/ Dixon del Tufo/O=Grid/O=UKHEP/ Cornwall/O=Grid/O=UKHEP/ Li/O=Grid/O=UKHEP/ D Mealor/O=Grid/O=UKHEP/ A Crosby/O=Grid/O=UKHEP/ Waters/O=Grid/O=UKHEP/ Cranfield/O=Grid/O=UKHEP/ West/O=Grid/O=UKHEP/ Walker/O=Grid/O=UKHEP/ Lewis/O=Grid/O=UKHEP/ D J Colling/O=Grid/O=UKHEP/ Howard /O=Grid/O=UKHEP/ Barlow/O=Grid/O=UKHEP/ Foster/O=Grid/O=UKHEP/ Forti/O=Grid/O=UKHEP/ Clarke/O=Grid/O=UKHEP/ Sansum /O=Grid/O=UKHEP/ Gordon/O=Grid/O=UKHEP/ McNab/O=Grid/O=UKHEP/ Hughes-Jones /O=Grid/O=UKHEP/ McCance/O=Grid/O=UKHEP/ Doyle /O=Grid/O=UKHEP/ /O=Grid/O=UKHEP/ Gordon

6 Tony Doyle - University of Glasgow GridPP Vision From Web to Grid - Building the next IT Revolution Premise The next IT revolution will be the Grid. The Grid is a practical solution to the data-intensive problems that must be overcome if the computing needs of many scientific communities and industry are to be fulfilled over the next decade. Aim The GridPP Collaboration aims to develop and deploy a large-scale science Grid in the UK for use by the worldwide particle physics community. Many Challenges.. Shared distributed infrastructure For all applications

7 Tony Doyle - University of Glasgow GridPP Objectives 1.SCALE: GridPP will deploy open source Grid software (middleware) and hardware infrastructure to enable the testing of a prototype of the Grid for the LHC of significant scale. 2.INTEGRATION: The GridPP project is designed to integrate with the existing Particle Physics programme within the UK, thus enabling early deployment and full testing of Grid technology and efficient use of limited resources. 3.DISSEMINATION: The project will disseminate the GridPP deliverables in the multi-disciplinary e-science environment and will seek to build collaborations with emerging non- PPARC Grid activities both nationally and internationally. 4.UK PHYSICS ANALYSES (LHC): The main aim is to provide a computing environment for the UK Particle Physics Community capable of meeting the challenges posed by the unprecedented data requirements of the LHC experiments. 5.UK PHYSICS ANALYSES (OTHER): The process of creating and testing the computing environment for the LHC will naturally provide for the needs of the current generation of highly data intensive Particle Physics experiments: these will provide a live test environment for GridPP research and development. 6.DATAGRID: Open source Grid technology is the framework used to develop this capability. Key components will be developed as part of the EU DataGrid project and elsewhere. 7.LHC COMPUTING GRID: The collaboration builds on the strong computing traditions of the UK at CERN. The CERN working groups will make a major contribution to the LCG research and development programme. 8.INTEROPERABILITY: The proposal is also integrated with developments from elsewhere in order to ensure the development of a common set of principles, protocols and standards that can support a wide range of applications. 9.INFRASTRUCTURE: Provision is made for facilities at CERN (Tier-0), RAL (Tier-1) and use of up to four Regional Centres (Tier-2). 10.OTHER FUNDING: These centres will provide a focus for dissemination to the academic and commercial sector and are expected to attract funds from elsewhere such that the full programme can be realised.

8 Tony Doyle - University of Glasgow GridPP Project Map - Elements

9 Tony Doyle - University of Glasgow Rare Phenomena – Huge Background 9 orders of magnitude! The HIGGS All interactions

10 Tony Doyle - University of Glasgow LHC Computing Challenge Tier2 Centre ~1 TIPS Online System Offline Farm ~20 TIPS CERN Computer Centre >20 TIPS RAL Regional Centre US Regional Centre French Regional Centre Italian Regional Centre Institute Institute ~0.25TIPS Workstations ~100 MBytes/sec Mbits/sec One bunch crossing per 25 ns 100 triggers per second Each event is ~1 Mbyte Physicists work on analysis channels Each institute has ~10 physicists working on one or more channels Data for these channels should be cached by the institute server Physics data cache ~PBytes/sec ~ Gbits/sec or Air Freight Tier2 Centre ~1 TIPS ~Gbits/sec Tier 0 Tier 1 Tier 3 Tier 4 1 TIPS = 25,000 SpecInt95 PC (1999) = ~15 SpecInt95 ScotGRID++ ~1 TIPS Tier 2

11 Tony Doyle - University of Glasgow Tier-0 - CERN Commodity Processors +IBM (mirrored) EIDE Disks Scale: ~1,000 CPUs ~5 PBytes Compute Element (CE) Storage Element (SE) User Interface (UI) Information Node (IN) Storage Systems..

12 Tony Doyle - University of Glasgow UK Tier-1 RAL New Computing Farm 4 racks holding 156 dual 1.4GHz Pentium III cpus. Each box has 1GB of memory, a 40GB internal disk and 100Mb ethernet. 50TByte disk-based Mass Storage Unit after RAID 5 overhead. PCs are clustered on network switches with up to 8x1000Mb ethernet out of each rack. Tape Robot upgraded last year uses 60GB STK 9940 tapes 45TB currrent capacity could hold 330TB Scale: 1000 CPUs 0.5 PBytes

13 Tony Doyle - University of Glasgow Regional Centres SRIF Infrastructure Local Perspective: Consolidate Research Computing Optimisation of Number of Nodes? 4 Relative size dependent on funding dynamics Global Perspective: V. Basic Grid Skeleton

14 Tony Doyle - University of Glasgow UK Tier-2 ScotGRID ScotGrid Processing nodes at Glasgow 59 IBM X Series 330 dual 1 GHz Pentium III with 2GB memory 2 IBM X Series 340 dual 1 GHz Pentium III with 2GB memory and dual ethernet 3 IBM X Series 340 dual 1 GHz Pentium III with 2GB memory and Mbit/s ethernet 1TB disk LTO/Ultrium Tape Library Cisco ethernet switches ScotGrid Storage at Edinburgh IBM X Series 370 PIII Xeon with 512 MB memory 32 x 512 MB RAM 70 x 73.4 GB IBM FC Hot- Swap HDD CDF equipment at Glasgow 8 x 700 MHz Xeon IBM xSeries GB memory 1 TB disk Griddev testrig at Glasgow 4 x 233 MHz Pentium II 2004 Scale: 300 CPUs 0.1 PBytes BaBar UltraGrid System at Edinburgh 4 UltraSparc 80 machines in a rack 450 MHz CPUs in each 4Mb cache, 1 GB memory Fast Ethernet and Myrinet switching

15 Tony Doyle - University of Glasgow Network Network Internal networking is currently a hybrid of –100Mb(ps) to nodes of cpu farms –1Gb to disk servers –1Gb to tape servers UK: academic network SuperJANET4 –2.5Gb backbone upgrading to 20Gb in 2003 EU: SJ4 has 2.5Gb interconnect to Geant US: New 2.5Gb link to ESnet and Abilene for researchers UK involved in networking development –internal with Cisco on QoS –external with DataTAG

16 Tony Doyle - University of Glasgow Grid issues – Coordination Technical part is not the only problem Sociological problems? resource sharing –Short-term productivity loss but long-term gain Key? communication/coordination between people/centres/countries –This kind of world-wide close coordination across multi-national collaborations has never been done in the past We need mechanisms here to make sure that all centres are part of a global planning –In spite of different conditions of funding, internal planning, timescales etc The Grid organisation mechanisms should be complementary and not parallel or conflicting to existing experiment organisation –LCG-DataGRID-eSC-GridPP –BaBar-CDF-D0-ALICE-ATLAS-CMS-LHCb-UKQCD Local Perspective: build upon existing strong PP links in the UK to build a single Grid for all experiments

17 Tony Doyle - University of Glasgow Experiment Deployment

18 Tony Doyle - University of Glasgow DataGrid Middleware Work Packages Collect requirements for middleware –Take into account requirements from application groups Survey current technology –For all middleware Core Services testbed –Testbed 0: Globus (no EDG middleware) First Grid testbed release Testbed 1: first release of EDG middleware WP1: workload –Job resource specification & scheduling WP2: data management –Data access, migration & replication WP3: grid monitoring services –Monitoring infrastructure, directories & presentation tools WP4: fabric management –Framework for fabric configuration management & automatic sw installation WP5: mass storage management –Common interface for Mass Storage Sys. WP7: network services –Network services and monitoring

19 Tony Doyle - University of Glasgow DataGrid Architecture Collective Services Information & Monitoring Replica Manager Grid Scheduler Local Application Local Database Underlying Grid Services Computing Element Services Authorization Authentication and Accounting Replica Catalog Storage Element Services SQL Database Services Fabric services Configuration Management Configuration Management Node Installation & Management Node Installation & Management Monitoring and Fault Tolerance Monitoring and Fault Tolerance Resource Management Fabric Storage Management Fabric Storage Management Grid Fabric Local Computing Grid Grid Application Layer Data Management Job Management Metadata Management Object to File Mapping Service Index

20 Tony Doyle - University of GlasgowAuthentication/Authorization Authentication (CA Working Group) –11 national certification authorities –policies & procedures mutual trust –users identified by CAs certificates Authorization (Authorization Working Group) –Based on Virtual Organizations (VO). –Management tools for LDAP-based membership lists. –6+1 Virtual Organizations VOs ALICEEarth Obs. ATLASBiomedical CMS LHCbGuidelines CAs CERN CESNET CNRS DataGrid- ES GridPP Grid-Ireland INFN LIP NIKHEF NorduGrid Russian DataGrid

21 Tony Doyle - University of Glasgow WP7 - EDG Authorisation grid-mapfile generation o=testbed, dc=eu-datagrid, dc=org CN=Franz Elmer ou=People CN=John Smith mkgridmap grid-mapfile VO Directory Authorization Directory CN=Mario Rossi o=xyz, dc=eu-datagrid, dc=org CN=Franz ElmerCN=John Smith Authentication Certificate ou=Peopleou=Testbed1ou=??? local usersban list

22 Tony Doyle - University of Glasgow Current User Base Grid Support Centre GridPP (UKHEP) CA uses primitive technology –It works but takes effort –201 personal certs issued –119 other certs issued GSC will run a CA for UK escience CA –Uses openCA; Registration Authority uses web –We plan to use it –Namespace identifies RA, not Project –Authentication not Authorisation Through GSC we have access to skills of CLRC eSC Use helpdesk to formalise support later in the rollout UK e-Science UK e-Science Certification Certification Authority Authority Scale Scale

23 Tony Doyle - University of Glasgow EDG TestBed 1 Status 30 Aug :38 Web interface showing status of (~400) servers at testbed 1 sites Production Centres

24 Tony Doyle - University of Glasgow GridPP Context (Externally) Neil Geddes Interoperability

25 Tony Doyle - University of Glasgow Interoperability Trust Relationships

26 Tony Doyle - University of Glasgow GridPP Sites in Testbed(s)

27 Tony Doyle - University of Glasgow GridPP Sites in Testbed: Status 30 Aug :38

28 Tony Doyle - University of Glasgow t0t0 t1t1 From Grid to Web… using GridSite

29 Tony Doyle - University of GlasgowDocumentation GridPP Web Site: EDG User Guide: EDG User Guide: A biomedical user point of view. JDL Howto: Document.pdf Document.pdf GDMP Guide:

30 Tony Doyle - University of Glasgow Job Submission 1. Authentication grid-proxy-init 2. Job submission to DataGrid dg-job-submit 3. Monitoring and control dg-job-status dg-job-cancel dg-job-get-output 4. Data publication and replication globus-url-copy, GDMP 5. Resource scheduling JDL, sandboxes, storage elements Linux text interfaces implemented GUIs next..

31 Tony Doyle - University of Glasgow Job Submission Example dg-job-submit /home/evh/sicb/sicb/bbincl jdl -o /home/evh/logsub/ bbincl jdl: # Executable = "script_prod"; Arguments = " ,v235r4dst,v233r2"; StdOutput = "file output"; StdError = "file err"; InputSandbox = {"/home/evhtbed/scripts/x509up_u149","/home/evhtbed/sicb/mcsend", "/home/evhtbed/sicb/fsize","/home/evhtbed/sicb/cdispose.class","/ home/evhtbed/v235r4dst.tar.gz","/home/evhtbed/sicb/sicb/bbincl sh","/home/evhtbed/script_prod","/home/evhtbed/sicb/sicb dat","/home/evhtbed/sicb/sicb dat","/home/evhtbed/sicb /sicb dat","/home/evhtbed/v233r2.tar.gz"}; OutputSandbox = {"job txt","D ","file output","file er r","job txt","job txt"};

32 Tony Doyle - University of Glasgow GUI - today

33 Tony Doyle - University of Glasgow GUI Future? Web Services Access via Grid Certificate

34 Tony Doyle - University of Glasgow GridPP – Achievements and Issues 1st Year Achievements Complete Project Map –Applications: Middleware: Hardware Fully integrated with EU DataGrid and LCG Projects Rapid middleware deployment /testing Integrated US-EU applications development e.g. BaBar+EDG Roll-out document for all sites in the UK (Core Sites, Friendly Testers, User Only). Testbed up and running at 15 sites in the UK Tier-1 Deployment 200 GridPP Certificates issued First significant use of Grid by an external user (LISA simulations) in May 2002 Web page development (GridSite) Issues for Year 2 Status: 19 Jul :52 GMT – keep monitoring and improve testbed deployment efficiency Importance of EU-wide development of middleware Integrated Testbed for use/testing by all applications Reduce integration layer between middleware and application software Integrated US-EU applications development Tier-1 Grid Production Mode Tier-2 Definitions and Deployment Integrated Tier-1 + Tier-2 Testbed Transfer to UK e-Science CA Integration with other UK projects e.g. AstroGrid, MyGrid…

35 Tony Doyle - University of Glasgow GridPP Sites in Testbed: Status 19 Jul :52 Project Map Software releases at each site

36 Tony Doyle - University of Glasgow GridPP – An Operational Grid From Web to Grid… Fit into UK e-Science structures LHC Computing – Particle physicists will use experience in distributed computing to build and exploit the Grid Infrastructure – tiered computing down to the physicist desktop Importance of networking Existing experiments have immediate requirements Non-technical issues = recognising/defining roles (at various levels) UK GridPP started 1/9/01 EU DataGrid First Middleware ~1/9/01 Development requires a testbed with feedback –Operational Grid Status: 25 Jun :38:47 GMT – a day in the life.. GridPP Testbed is relatively small scale – migration plans reqd. e.g. for CA. Grid jobs are being submitted today.. user feedback loop is important.. Grid tools web page development by a VO. Next stop. Web services…

37 Tony Doyle - University of GlasgowSummary A vision is only useful if its shared Grid success is fundamental for PP 1.Scale in UK? 0.5 Pbytes and 2,000 distrib. CPUs GridPP in Sept Integration – ongoing.. 3.Dissemination – external and internal 4.LHC Analyses – ongoing feedback mechanism.. 5.Other Analyses – closely integrated using EDG tools 6.DataGrid - major investment = must be (and is so far) successful 7.LCG – Grid as a Service 8.Interoperability – sticky subject 9.Infrastructure – Tier-A/1 in place, Tier-2s to follow… 10.Finances – (very well) under control Next steps on framework VI.. CERN = EUs e-science centre? Co-operation required with other disciplines/industry esp. AstroGrid

Download ppt "Tony Doyle GridPP – Project Elements UK e-Science All Hands Conference, Sheffield 3 September 2002."

Similar presentations

Ads by Google