les robertson cernit 1

1
LHC Computing Grid Project - LCG

• Project Status Summary
• POB - 30 August 2004
• Les Robertson LCG Project Leader
• CERN European Organization for Nuclear Research
• Geneva, Switzerland
• les.robertson_at_cern.ch

2
Key Points (i)

• Grid Deployment
• Very many sites have joined, with more nominal
  capacity than we expected at this time
• Each experiment has full control over the sites
  that it accepts to use
• The accounting system is now in the certification
  phase
• There has been continued progress in
  inter-operation with other grid infrastructures
  Grid3 and NorduGrid
• First service challenge starting
• The Grid Deployment steering group has been
  established to oversee
• Service challenges
• Inter-operability
• Operational problem escalation

3
Key Points (ii)

• Grid Deployment has entered a new phase
• Basic middleware is basically working
• responsible now for a small fraction of the
  problems
• Many performance issues
• some solutions are (slowly) being developed
• points to priorities for gLite team
• Many operational issues
• mis-configuration, out of date middleware
• resources unsuitable for experiments needs (e.g.
  insufficient disk space)
• slow response by sites to problems exacerbated
  by the holiday period, security concerns
• We are a long way from the grid appearing as a
  single coherent facility experiments must adapt
  to the current state of development
• And the result has been very effective for LHCb -
  3000 concurrent jobs last Friday

4
Key Points (iii)

• Fabric
• Agreement confirmed on Scientific Linux
• Security issues an increasing problem and will
  lead to increased pressure for restrictions on
  network visibility
• The trend towards small files requires urgent
  discussion
• Applications
• Good collaboration on improving the integration
  of ROOT and the other applications area tools
  but the test will be the common maths library
• The conditions database project is progressing
  good statements of support from experiments but
  so far there are no commitments to provide
  manpower essential to the continuation of the
  project
• Gabriele Cosmo replaces Torre Wenaus as the
  Simulation Project Leader

5
Key Points (iv)

• ARDA
• The ARDA project is now staffed and has agreed a
  detailed programme of work with three of the
  experiments
• First experiences with the initial gLite
  prototype were discussed at the ARDA workshop in
  June next workshop scheduled for October, when
  there will also be first experience with the
  experiments analysis facilities
• Middleware
• Agreement on the split of responsibilities
• Grid Deployment Area short term developments
  driven by operational priorities
• gLite team responsible for long term developments
• with external compatibility wherever possible
• First gLite prototype in mid-May
• Architecture document
• Co-existence with LCG-2 single pre-production
  testbed

6
Key Points (v)

• Resources
• Phase 2 Planning group has collected Phase 2
  plans from most of the major centres
• Some difficulty now in getting probable/possible
  breakdown of resources between experiments
  essential to complete the planning round
• Significant shortfall in planned disk resources
• Question of the process that is required before a
  site can take account of the updated experiment
  requirements

7
Key Points (vi)

• Resources at CERN for Phase 2
• In the Council paper on cost of completion (March
  2002) the funding foreseen in the CERN budget
  fell short of the estimated costs by CHF 20M
• The latest planning round for materials and staff
  stays within the original cost estimates
• But the deficit with respect to the funding
  committed at CERN is still CHF 20M
• The DG proposed in June Council that funding
  agencies look at the possibilities for
  continued/new special funding for staff at CERN
  in Phase 2. So far only Spain has responded to
  this call.
• Because of the extension of Phase 1 from 3 to 4
  years (delay of LHCC startup) the staff profile
  (heavily dependent on special funding) falls off
  rapidly from early 2005
• Key staff will shortly receive termination
  notices
• Therefore it is essential to decide on how to
  deal with the funding shortfall in the next months

8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
The details would haveto be discussed thisis
an example to showthe depth of cutting
thatwould be required
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
Advice needed

• There is some probability that there will be
  additional funding
• the solution proposed by the DG for special
  funding for staff
• moving the cost of tape media to the funding
  agencies, as was always done for previous
  experiments
• calling on the CERN contingency as considered
  at the time of the cost to completion paper
• A shortfall of CHF 10M would have a significant
  effect on the materials budget increasing the
  risk that we are not prepared for LHC data
  analysis, and requiring additional analysis
  capacity outside CERN.
• The project leader does not believe that the
  costs can be cut any further without a clear
  reduction in the scope of the CERN part of the
  project
• But the two solutions proposed to reduce the role
  of the host lab require outside centres to take
  on very significant responsibilities.
• We need to find a solution within three months
  that enables us to plan the Phase 2 staffing at
  CERN