Beyond GridPP2 and e-infrastructure.

Panel:

DB:  Issues: Funding, hardware, support, Tier1/Tier2 relationship,
     GridPP/EGEE relationship.

DK:  Timing of end of GridPP2 bad.  Assumption will be finished at end
     of GridPP2 but won't be:
      - Operation of production service still not tackled fully.
      - How to do physics analysis.
     Aim to maximise physics output.  Need for 10 million pounds a year
     a challenge - balance Grid/Physics.
     Can't run a service with limited people - need ongoing development.
     Good in UK, groups working well together now, how do we keep that
     going?

SL:  Can meet Tier2 requirements for next few years.  Hardware support
     uncertain funding yet, will need more.  Need to keep up support
     for Tier2s.

TD:  e-science - what does it mean?  Enable science and infrastructure?
     Planning starting now in PPARC.  By May will know gross allocation.
     Need to sustain large EU project involvement.  All costs have large
     uncertainties.

JG:  EG project 2nd phase April 2006 to 2008.  Bringing together grid
     projects in EU.
     Two grids in UK - GridPP and National Grid Service - need to bring
     them together, also Grid Ireland.
     We should try to use NGS to try to break down boundaries.
     Similar across other research councils - no more special projects -
     grid part of general funding.  Cooperate on networking, security,
     single UK infrastructure, not PPARC funding.

AD:  Middleware - still requirements not satisfied.
     Need to move to web service standards.

DN:  For experiments would like infrastructure posts after 2007.
     Need to keep spirit of collaboration not just hand money to the
     experiments.

Discussion:

     A few people have run jobs on NGS.
     Grid should be used for other projects - more generic.
     Rebrand as National Grid Service?
     We don't have enough resources to give much away, but knowledge is
     an important thing to share.
     GridPP achieved a lot by being involved internationally - not true
     for NGS.
     Must stay focussed on particle physics.
     Analogy with networking starting within the community then filtering
     outside and eventually becoming completely separate.
     Situation varies in other countries - some have multiple grids like
     us, others just one.
     What is the grid - for most users a glorified batch system,
     submitting jobs and moving data.
     Command line commands need to be stable for people like Pheno who
     don't have the resources to maintain their own front ends.
     Tight time scale for us but there won't be another funded GridPP3.
     CA now out of our hands, other parts will be over time.
     Keep in mind that things provided by GridPP have passed to the
     wider community.