A common LHC approach to GRID F Carminati

A common LHC approach to GRID F. Carminati HICB – Toronto February 17, 2002 1

Current situation The LHC experiments collaborates within Data. GRID-WP 8 A lot of ground has been covered toward a constructive dialog and coordination – however n n We still are building a vision of what we need from a GRID We still do not have a well defined, high-level, GRID architecture to adapt our applications to Our long term use cases still lack coherence n n In reality we are doing more or less the same thing! We produce read only files with atomic entities (events) We refine them progressively We do data mining February 17, 2002 HICB 2

What we want from a GRID Specific application layer VO common application layer GRID architecture GLOBUS team ALICE ATLAS CMS LHCb LHC Other apps High level GRID middleware PPDG, Gri. Phyn, Data. GRID, i. VDGL Basic Services OS & Net services February 17, 2002 HICB 3

What we have Specific application layer ALICE ATLAS MW 1 CMS MW 2 LHCb MW 3 MW 4 Other apps MW 5 Middle. Ware GLOBUS team Basic Services OS & Net services February 17, 2002 HICB 4

How to proceed CMS ATLAS Core common use case ALICE February 17, 2002 LHCb HICB 5

A proposal Orwe even It will bebetter easiertotodefine arrive at If manage Specific application layer VO use cases & requirements ALICE ATLAS CMS LHCb Other apps LHC Other apps Common core case Common useuse cases MW 1 MW 2 MW 3 MW 4 MW 5 Middle. Ware GLOBUS team Bag of Services (GLOBUS) OS & Net services February 17, 2002 HICB 6

A proposal Extend dialogue beyond the scope of Data. GRID to the GRID projects of the LHC experiments at the LCG level n This includes and indeed demands EU-US coordination for ALICE, ATLAS and CMS We should build on n n The current dialogue established within WP 8 The testbed plans we are discussing since more than one year in Data. GRID But the result should be MW-neutral A common set of use-cases would help n n Clarify and rationalize our demands to the different GRID MW projects (i. VDGL, Gri. Phy. N, PPDG, EDG, …) Run trans-Atlantic tesbeds making meaningful comparisons February 17, 2002 HICB 7

GRID interoperation Application Gri. Phy. N – PPDG – i. VDGL EDG February 17, 2002 HICB 8

Why this is important? Experiments want to work on common LCG projects n We need a common set of requirements / use cases to define common deliverables The HICB expects proposal from the experiments n n n It would be effective to provide a common set of use cases Instead of competing one with the other Of course current proposals should go ahead as they will provide useful input to this exercise The different GRID MW activities risk to diverge n n Common use cases could help them to develop coherent solutions Or ideally complementary elements February 17, 2002 HICB 9

How to proceed LCG SC 2 approved a Requirement Technical Assessment Group (RTAG) on common GRID use cases n This RTAG should work out an initial set of high-level use cases and requirements, and indicate how those should be refined Current plans should not be hindered n n Indeed they should be encouraged as they provide material for the discussion The experience of EDG TB 1 has been very positive in this sense New tests implementing the common use case should be developed by the experiments and progressively replace or complement the current plans February 17, 2002 HICB 10

RTAG for common GRID use cases & requirements Chair: F. Carminati Mandate: define a common set of use cases and requirements for GRID applications among the LHC experiments n n Review the current plans of integration of GRID services in the software frameworks of the experiments Identify and describe a set of high level use cases of GRID technology common to the four experiments Identify and describe which use cases will be specific for the different experiments Derive a set of common requirements for GRID MW Timescale - Initial report in 1 month, final report within 2 months Makeup - Reps from experiments, EDG, i. VDGL + additional members at chair’s discretion Guidance – The result of the RTAG should be the detailed description of use cases implementable by the framework of the four experiments using present or planned GRID middleware. These should serve to the MW developers (both in US and in Europe) to guide their work and to the experiments as platform to perform GRID interoperability studies February 17, 2002 HICB 11