infso-ri-508833 enabling grids for e-science egee – application support and identification...

19
INFSO-RI-508833 Enabling Grids for E-sciencE www.eu-egee.org EGEE – application support and identification Vincent Breton, on behalf of NA4 Application identification and support manager CNRS, France Second EGEE review December 6th 2005

Upload: clarence-todd

Post on 29-Jan-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

  • Metadata: ARDA Implementation

    PrototypeValidate our ideas and expose a concrete example to interested partiesMultiple back endsCurrently: Oracle, PostgreSQL, SQLite, MySQLDual front endsTCP Streaming Chosen for performanceSOAPFormal requirement of EGEECompare SOAP with TCP StreamingAlso implemented as standalone Python libraryData stored on the file system

    Massimo Lamanna / CERN

    Python Interpreter

    Metadata Python API

    Client

    filesystem

    Database

    MDServer

    Metadata Server

    SOAP

    TCP Streaming

    Oracle

    PostgreSQL

    SQLite

    Client

    Client

    INFSO-RI-508833

    NA4 biomedical gLite tests on the PPSPPS components tested: Weekly testing of WMS, FiReMan, security related components (ACLs for file access), R-GMAOther gLite testingSite installation (v1.0, 1.1, 1.2 and 1.3)Advanced data management on prototype testbedMedical ImagerTrigger:Retrieve DICOM files from imager.Register file in FiremangLite EDS client for data encryptionRegister Metadata in AMGAClient Library:Lookup file through Metadata (AMGA)Use gLite EDS client for data decryption Serve it up to the application

    INFSO-RI-508833

    All gLite Services are available and tested on GILDAgLite Services

    INFSO-RI-508833

    2. The capture of requirements, and the assessment of middlewareThe status at PM18

    A database of requirements has been constructed by the Project Technical Forum(PTF)This database already contains >400 requirementsThis database is populated by the scientific communities deploying applications on EGEE and by the FP6 projects (eg Diligent, Seegrid,...)

    The capture of requirements from FP5 projects

    Several FP5 projects were granted access to the database (Grace, Mammogrid)But it has proved to be difficult to collect their requirementsNo formalization of requirementsProjects finishedRequirements were collected mainly from the scientific communities (Crossgrid, Datagrid)

    The assessment of middleware w.r.t requirements

    There is on-going work with JRA1 within the framework of the PTF and the TCG

    INFSO-RI-508833

    3. The clarifification of the true motivation of usersAn improved Memorandum of Understanding(MOU) between EGEE and the scientific communities has been designed and implemented

    Detailed SA1 questionnaire to evaluate the required resourcesRegular phone conferences with the Operation Advisory Group

    An improved application integration process is under design

    Creation of lightweight recognised VOs vs supported Vos

    Only supported VOs will be asked to write an MOU

    Decentralised integration vs centralised EGAAP-like approach

    Initial deployment will be at regional level

    INFSO-RI-508833

    The status of the MoUs with the generic supported applications

    INFSO-RI-508833

    The status of generic applications deployment on EGEEApplications accepted by the Project Executive Board(PEB)

    Earth Science Research (Earth Observation, Hydrology, Climate)Geophysics (Industry)Computational ChemistryAstrophysics (MAGIC and Planck collaborations) Finance (EGRID)

    New Applications recommended to the PEB by the EGEE Generic Applications Advisory Panel (EGAAP)

    Fusion (ITER)ArchaeologyEC projects (EELA, EUMEDGRID, EUCHINAGRID, BIOINFOGRID)

    INFSO-RI-508833

    4. The evaluation of user satisfactionFirst user survey at PM1534 users filled a questionnaire available onlineResults available in DNA4.3.2User satisfaction is good for user supportUser satisfaction on grid services is intermediateUsers would like improvements in documentation (with the exception of system manuals which are well appreciated)Action: creation of the User information Group to address issues related to documentation in the project Next user survey at PM21

    INFSO-RI-508833

    5. Demonstration of the scientific benefits of the EGEE gridThe number of users in the VOs related to NA4 activity has continued to grow

    from ~500 at PM9 to ~1000 at PM18 More than 20 applications are deployed on the production infrastructure

    WE HAVE DEMONSTRATED

    added value in terms of deployment scale

    The very large scale production of simulated data for HEP experiments

    In particular this has been essential for the commissioning of the LHC experiments

    added value in terms of response time

    The computation of the seismic epicentre by ESRThe search for new drugs: large scale in silico docking (WISDOM) - access to large scale resources gives much improved turnaround

    added value in terms of data distribution

    The distribution of LHC data on the grid, and the large users activity (~150k jobs/month in the CMS case)

    INFSO-RI-508833

    Large Scale Production on EGEE by several HEP experiments Fundamental activity in the preparation of the LHC experiments start up

    Physics and detector studiesComputing systems commissioning

    Examples:

    ATLAS: Ran on 3 grids (2/3 of work on EGEE) Peaks of 10,000 jobs per day on EGEE200 CPU years in 2005 ; 45 TB dataComprehensive analysis: see S.Campana et al., e-Science 2005, Melbourne, Australia

    LHCb:Peaks of 5,500 jobs/day on EGEE700 CPU/years in 2005 ; 77 TB of data

    A lot of activity within non-LHC experiments (BaBar,CDF, D, ZEUS,H1)

    More details in DNA4.3.2ATLAS production evolution 2004-2005

    INFSO-RI-508833

    First biomedical data challenge: World-wide In Silico Docking On Malaria (WISDOM)

    Significant biological parameterstwo different molecular docking applications (Autodock and FlexX)about one million virtual ligands selectedtarget proteins from the parasite responsible for malariaSignificant numbers Total of about 46 million ligands docked in 6 weeks1TB of data produced Up 1000 computers in 15 countries used simultaneously corresponding to about 80 CPU yearsAverage crunching factor ~600WISDOM open dayDecember 16th, 2005, Bonn (Germany)

    Discuss Data Challenge resultsPrepare next steps towards a malaria Grid (EGEE-II, Embrace, Bioinfogrid)Information: http://wisdom.eu-egee.fr

    Graph2

    597

    1072

    383

    715

    5122

    7580

    2004

    263

    3687

    3356

    1877

    337

    218

    827

    8106

    Domain distribution of Flexx run jobs

    flexx

    flexx

    479

    62

    56

    659

    55

    283

    1072

    662

    129

    151

    383

    1877

    189

    56

    1759

    766

    874

    1014

    811

    612

    2488

    2600

    1361

    382

    133

    74

    460

    97

    125

    282

    644

    9

    102

    2690

    1347

    4940

    3

    337

    666

    715

    56

    263

    827

    380

    2763

    218

    233

    nb of run jobs

    CE distribution for Flexx run jobs

    autodock

    479

    62

    56

    659

    55

    283

    1072

    662

    129

    151

    383

    1877

    189

    56

    1759

    766

    874

    1014

    811

    612

    5088

    1361

    382

    133

    74

    460

    97

    125

    282

    644

    9

    102

    2690

    1347

    4940

    3

    337

    666

    715

    56

    263

    827

    380

    2763

    218

    233

    Site distribution for DC run jobs (Flexx)

    paysFlexx

    paysFlexx

    229

    14

    286

    9

    17

    174

    301

    603

    757

    141

    230

    1082

    143

    13

    699

    108

    418

    708

    436

    255

    912

    1136

    445

    201

    483

    5

    45

    107

    9

    324

    165

    59

    703

    1080

    585

    1716

    414

    119

    470

    128

    35

    250

    465

    221

    854

    17

    83

    842

    nb of run jobs

    CE distribution for Autodock run jobs

    paysautodock

    229

    14

    286

    9

    17

    174

    301

    603

    757

    141

    230

    1082

    143

    13

    699

    108

    418

    708

    436

    255

    2048

    445

    201

    483

    5

    45

    107

    9

    324

    165

    59

    703

    1080

    585

    1716

    414

    119

    470

    128

    35

    250

    465

    221

    854

    17

    83

    842

    Site distribution of DC Biomed run jobs (Autodock)

    data

    paysnb

    bg479

    bg62

    bg56

    Total bg597bg597

    com1072com1072

    Total com1072com1072

    cy383cy383

    Total cy383cy383

    de715de715

    Total de715de715

    es283es283

    es662es662

    es151es151

    es874es874

    es9es9

    es380es380

    es2763es2763

    Total es5122es5122

    fr55fr55

    fr1014fr1014

    fr811fr811

    fr612fr612

    fr2488fr2488

    fr2600fr2600

    Total fr7580fr7580

    gr189gr189

    gr56gr56

    gr1759gr1759

    Total gr2004gr2004

    il263il263

    Total il263il263

    it129it129

    it1361it1361

    it382it382

    it133it133

    it74it74

    it460it460

    it97it97

    it125it125

    it282it282

    it644it644

    Total it3687it3687

    nl2690nl2690

    nl666nl666

    Total nl3356nl3356

    pl1877pl1877

    Total pl1877pl1877

    ro337ro337

    Total ro337ro337

    ru218ru218

    Total ru218ru218

    tw827tw827

    Total tw827tw827

    uk659uk659

    uk766uk766

    uk102uk102

    uk1347uk1347

    uk4940uk4940

    uk3uk3

    uk56uk56

    uk233uk233

    Total uk8106uk8106

    Total36144

    data

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    Domain distribution of Flexx run jobs

    AUTODOCKpaysnb

    bg229

    bg14

    Total bg243bg243

    com301com301

    Total com301com301

    cy230cy230

    Total cy230cy230

    de128de128

    Total de128de128

    es174es174

    es603es603

    es141es141

    es418es418

    es59es59

    es221es221

    es854es854

    Total es2470es2470

    fr17fr17

    fr708fr708

    fr436fr436

    fr255fr255

    fr912fr912

    fr1136fr1136

    Total fr3464fr3464

    gr143gr143

    gr13gr13

    gr699gr699

    Total gr855gr855

    il250il250

    Total il250il250

    it757it757

    it445it445

    it201it201

    it483it483

    it5it5

    it45it45

    it107it107

    it9it9

    it324it324

    it165it165

    Total it2541it2541

    nl1080nl1080

    nl470nl470

    Total nl1550nl1550

    pl1082pl1082

    Total pl1082pl1082

    ro119ro119

    Total ro119ro119

    ru17ru17

    ru83ru83

    Total ru100ru100

    tw465tw465

    Total tw465tw465

    uk286uk286

    uk9uk9

    uk108uk108

    uk703uk703

    uk585uk585

    uk1716uk1716

    uk414uk414

    uk35uk35

    uk842uk842

    Total uk4698uk4698

    Total18496

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    0

    Domain distribution of Autodock run jobs

    SiteCEdomainrun jobs

    -1BG01-IPPce001.grid.bas.bgbg229AUTODOCKBG01-IPP229

    -1BG02-IMce001.imbm.bas.bgbg14BG02-IM14

    -1BHAM-LCG2epgce1.ph.bham.ac.ukuk286BHAM-LCG2286

    -1CAVENDISH-LCG2serv03.hep.phy.cam.ac.ukuk9CAVENDISH-LCG29

    -1CEA-DAPNIA-SACLAYnode03.datagrid.cea.frfr17CEA-DAPNIA-SACLAY17

    -1CESGA-EGEEce2.egee.cesga.eses174CESGA-EGEE174

    -1CGG-LCG2ce1.egee.fr.cgg.comcom301CGG-LCG2301

    -1CIEMAT-LCG2lcg02.ciemat.eses603CIEMAT-LCG2603

    -1CNAF-T1ce01-lcg.cr.cnaf.infn.itit757CNAF-T1757

    -1CNB-LCG2mallarme.cnb.uam.eses141CNB-LCG2141

    -1CY01-LCG2ce101.grid.ucy.ac.cycy230CY01-LCG2230

    -1CYFRONET-LCG2zeus02.cyf-kr.edu.plpl1082CYFRONET-LCG21082

    -1GR-01-AUTHnode001.grid.auth.grgr143GR-01-AUTH143

    -1GR-04-FORTH-ICSgrid001.ics.forth.grgr13GR-04-FORTH-ICS13

    -1HG-01-GRNETce01.isabella.grnet.grgr699HG-01-GRNET699

    -1IC-LCG2gw39.hep.ph.ic.ac.ukuk108IC-LCG2108

    -1IFCA-LCG2egeece.ifca.org.eses418IFCA-LCG2418

    -1IN2P3-CCcclcgceli01.in2p3.frfr708IN2P3-CC708

    -1IN2P3-CPPMmarseillece01.mrs.grid.cnrs.frfr436IN2P3-CPPM436

    -1IN2P3-LALgrid10.lal.in2p3.frfr255IN2P3-LAL255

    -1IN2P3-LPCclrce01.in2p3.frfr912IN2P3-LPC2048

    -1IN2P3-LPCclrce02.in2p3.frfr1136INFN-BARI445

    -1INFN-BARIgridba2.ba.infn.itit445INFN-CAGLIARI201

    -1INFN-CAGLIARIgrid002.ca.infn.itit201INFN-CATANIA483

    -1INFN-CATANIAgrid012.ct.infn.itit483INFN-CNAF5

    -1INFN-CNAFgridit-ce-001.cnaf.infn.itit5INFN-FERRARA45

    -1INFN-FERRARAgrid0.fe.infn.itit45INFN-NAPOLI107

    -1INFN-NAPOLIgriditce01.na.infn.itit107INFN-PADOVA9

    -1INFN-PADOVAprod-ce-01.pd.infn.itit9INFN-ROMA1324

    -1INFN-ROMA1t2-ce-01.roma1.infn.itit324INFN-ROMA1-VIRGO165

    -1INFN-ROMA1-VIRGOvirgo-ce.roma1.infn.itit165INTA-CAB59

    -1INTA-CABce00.inta.eses59ManHEP-LCG2703

    -1ManHEP-LCG2bohr0001.tier2.hep.man.ac.ukuk703NIKHEF-ELPROD1080

    -1NIKHEF-ELPRODtbn20.nikhef.nlnl1080OXFORD-01-LCG2585

    -1OXFORD-01-LCG2t2ce02.physics.ox.ac.ukuk585RAL-LCG21716

    -1RAL-LCG2lcgce01.gridpp.rl.ac.ukuk1716RALPP-LCG414

    -1RALPP-LCGheplnx201.pp.rl.ac.ukuk414ROGRID-ICI119

    -1ROGRID-ICItestbed001.grid.ici.roro119SARA-LCG2470

    -1SARA-LCG2mu6.matrix.sara.nlnl470SCAI128

    -1SCAIscaicl0.scai.fraunhofer.dede128ScotGRID-Edinburgh35

    -1ScotGRID-Edinburghce.epcc.ed.ac.ukuk35TAU-LCG2250

    -1TAU-LCG2lcfgng.cs.tau.ac.ilil250Taiwan-LCG2465

    -1Taiwan-LCG2lcg00125.grid.sinica.edu.twtw465UPV-GRyCAP221

    -1UPV-GRyCAPramses.dsic.upv.eses221pic854

    -1picce01.pic.eses854ru-Moscow-SINP-LCG217

    -1ru-Moscow-SINP-LCG2lcg06.sinp.msu.ruru17ru-PSN-LCG283

    -1ru-PSN-LCG2lcgce.psn.ruru83scotgrid-gla842

    -1scotgrid-glace1-gla.scotgrid.ac.ukuk842

    1BG01-IPPce001.grid.bas.bgbg479FLEXXBG01-IPP479

    1BG02-IMce001.imbm.bas.bgbg62BG02-IM62

    1BG04-ACADce01.grid.acad.bgbg56BG04-ACAD56

    1BHAM-LCG2epgce1.ph.bham.ac.ukuk659BHAM-LCG2659

    1CEA-DAPNIA-SACLAYnode03.datagrid.cea.frfr55CEA-DAPNIA-SACLAY55

    1CESGA-EGEEce2.egee.cesga.eses283CESGA-EGEE283

    1CGG-LCG2ce1.egee.fr.cgg.comcom1072CGG-LCG21072

    1CIEMAT-LCG2lcg02.ciemat.eses662CIEMAT-LCG2662

    1CNAF-T1ce01-lcg.cr.cnaf.infn.itit129CNAF-T1129

    1CNB-LCG2mallarme.cnb.uam.eses151CNB-LCG2151

    1CY01-LCG2ce101.grid.ucy.ac.cycy383CY01-LCG2383

    1CYFRONET-LCG2zeus02.cyf-kr.edu.plpl1877CYFRONET-LCG21877

    1GR-01-AUTHnode001.grid.auth.grgr189GR-01-AUTH189

    1GR-04-FORTH-ICSgrid001.ics.forth.grgr56GR-04-FORTH-ICS56

    1HG-01-GRNETce01.isabella.grnet.grgr1759HG-01-GRNET1759

    1IC-LCG2gw39.hep.ph.ic.ac.ukuk766IC-LCG2766

    1IFCA-LCG2egeece.ifca.org.eses874IFCA-LCG2874

    1IN2P3-CCcclcgceli01.in2p3.frfr1014IN2P3-CC1014

    1IN2P3-CPPMmarseillece01.mrs.grid.cnrs.frfr811IN2P3-CPPM811

    1IN2P3-LALgrid10.lal.in2p3.frfr612IN2P3-LAL612

    1IN2P3-LPCclrce01.in2p3.frfr2488IN2P3-LPC50885088

    1IN2P3-LPCclrce02.in2p3.frfr2600INFN-BARI1361

    1INFN-BARIgridba2.ba.infn.itit1361INFN-CAGLIARI382

    1INFN-CAGLIARIgrid002.ca.infn.itit382INFN-CATANIA133

    1INFN-CATANIAgrid012.ct.infn.itit133INFN-CNAF74

    1INFN-CNAFgridit-ce-001.cnaf.infn.itit74INFN-FERRARA460

    1INFN-FERRARAgrid0.fe.infn.itit460INFN-NAPOLI97

    1INFN-NAPOLIgriditce01.na.infn.itit97INFN-PADOVA125

    1INFN-PADOVAprod-ce-01.pd.infn.itit125INFN-ROMA1282

    1INFN-ROMA1t2-ce-01.roma1.infn.itit282INFN-ROMA1-VIRGO644

    1INFN-ROMA1-VIRGOvirgo-ce.roma1.infn.itit644INTA-CAB9

    1INTA-CABce00.inta.eses9ManHEP-LCG2102

    1ManHEP-LCG2bohr0001.tier2.hep.man.ac.ukuk102NIKHEF-ELPROD2690

    1NIKHEF-ELPRODtbn20.nikhef.nlnl2690OXFORD-01-LCG21347

    1OXFORD-01-LCG2t2ce02.physics.ox.ac.ukuk1347RAL-LCG24940

    1RAL-LCG2lcgce01.gridpp.rl.ac.ukuk4940RALPP-LCG3

    1RALPP-LCGheplnx201.pp.rl.ac.ukuk3ROGRID-ICI337

    1ROGRID-ICItestbed001.grid.ici.roro337SARA-LCG2666

    1SARA-LCG2mu6.matrix.sara.nlnl666SCAI715

    1SCAIscaicl0.scai.fraunhofer.dede715ScotGRID-Edinburgh56

    1ScotGRID-Edinburghce.epcc.ed.ac.ukuk56TAU-LCG2263

    1TAU-LCG2lcfgng.cs.tau.ac.ilil263Taiwan-LCG2827

    1Taiwan-LCG2lcg00125.grid.sinica.edu.twtw827UPV-GRyCAP380

    1UPV-GRyCAPramses.dsic.upv.eses380pic2763

    1picce01.pic.eses2763ru-PSN-LCG2218

    1ru-PSN-LCG2lcgce.psn.ruru218scotgrid-gla233

    1scotgrid-glace1-gla.scotgrid.ac.ukuk233

    INFSO-RI-508833

    Example of a generic applications use of EGEE (ESR) OzonemapsEarthquakes epicentre determination

    INFSO-RI-508833

    Conclusions and a forward lookAchievements

    The number of users in the VOs related to NA4 activity has doubled

    The usage of the grid by the pilot applications has significantly evolved into large scale production

    EGAAP applications have also achieved production status on EGEE (ESR,MAGIC)

    The migration of several existing applications to gLite has been achieved

    A detailed MOU procedure has been defined and executed for several selected applications

    The first User Survey has been performed Issues being actively addressed

    Reduction of the amount of skilled effort required by the VOs for production use of the grid (EGEE is improving middleware, monitoring, user interfaces and user support)

    The availability, accessibility and quality of user documentation (an inter-activity group is working on this)

    Looking forwards : User Forum, EGEE-II

    INFSO-RI-508833

    EGEE first User ForumDates: March 1-3 2006Location: CERN, SwitzerlandTarget attendance: 150 participantsGoalsObtain a consistent understanding across the EGEE related projects of their expectations, present status and possible evolutionLearn how to improve EGEE quality of servicePromote cross-application fertilisationPrepare EGEE-IIParticipation open to external projects and EGEE membersFormat: 3-day workshopPresentations by thematic areas selected by invitation and through a call for contributionsEGEE presentations (integration of new applications, access to resources, status of middleware,...)With a lot of time for discussionURL:http://egee-intranet.web.cern.ch/egee-intranet/User-Forum/index.html

    Montrer un slide sur le user survey (evaluation of user satisfaction)Mieux expliciter la user documentation qui manque, Expliquer ce qui est en cours pour repondre au besoin

    ****ACLs for file access -> which componentChanger ce slide pour mettre les tests sur le PPSMake it explicit why egeode MoU is not finished Adding Diligent

    LHCb list of texts to eraseImprove resolution, too heavy, too many images, bad slide

    Expliquer pourquoi on a besoin Speak about EGEE-II