2
3
FromComputationalGrids
toe-Science
• TheGrid:complexnetworkedresources
– Don’tknowwhichresourceswillbeavailableandwhen
– Potentiallyfragileresourceconnectivityandavailability
– Linkinstruments/storage/computation/HPC
• E-Science:higher-levelapplieduseoftheseresources
– Complexapplicationshavetoretainperformance
• Analysisofverylargedistributeddatasets
• Coupledexecutionformulti-physicsorvisualisation
– Easeofusefordomainspecialists
• FortheDeveloper– higherlevelabstractions
• FortheUser– higherlevelinteraction,e.g.portals,PSE’s
– Betterutilisationofresourcesandimprovethroughput
– Supportscientificcollaborationandknowledgemanagement
4
BuildingComputational
Communities
• Needtofederateresourcesfromrealorganisations
– Expressownershipandretaincontrol
– Gridresourcesare‘not’free– accountability/payment!
– Managetheresourcesasasingleunit
• ComputationalCommunities(a.k.a.VirtualOrganisations)
– Shareandcombineheterogeneousresources
– Ensuretransparencybyhidingcomplexityfromtheuser
– Optimiseresourceutilisationforalljobsandusers
– Easytocontribute(andwithdraw)resources
• Collectandretaininformationrelatingto:
– Theresources
– Theapplications
– Theusers