MONARC%20-%20synopsis%20of%20recent%20activities

1
MONARC - synopsis of recent activities

• Planning regional centres
• Models and simulation

2
Site and Network Architecture - Documents

• http//nicewww.cern.ch/les/monarc/base_config.htm
  l
• Rough Sizing Estimates for a Minimal Regional
  Centre L Robertson
• http//home.cern.ch/barone/monarc/RCArchitecture.
  txt
• Regional Centers for LHC computing
  L Barone
• http//home.fnal.gov/odell/monarc_report.html
• Monarc Report on Existing Computing Architectures
  J Butler(FNAL) et al

3
Presentations from national RC contacts
(April 13)

• IN2P3/ France D. Linglin
• INFN / Italy F. Ruggieri
• LBNL / US (ATLAS) C. Tull
• Fermilab / US (CMS) M. Kasemann
• UK J. C. Gordon
• Germany A. Putzer
• KEK / Japan K. Amako
• Moscow / Russia V. Kolossov

4
Key points from national presentations

• IN2P3 D Linglin
• Currently 1700 SI95, 2 TB disk, 0.6 PB tape
• 40 computer specialists for supporting
  infrastructure
• Use ATMEthernet..going to Gigabit Ethernet
• IBMHP platforms.adding SUN(Babar) PC
  Linux(CDF)
• Linux better than NT for farms but not
  free!(human resource)
• Keep constant budgetfollow technology
• Await lHC requests from experiments
• INFN F Ruggieri
• Want a national centre open to all experiments,
  and with open h/w and s/w
• Probably go for one national centre ?? (2002
  decisionoperate end 2004)
• See raw data at CERN with Multi Gbit link(s) to
  CERN

5

• LBL( USA Atlas ) C Tull
• US Atlas 20 Atlas
• Raw data and reconstruction at CERN
• 200 TB DST data/yearal to go to RC
• 120 active users (40 active?)
• 2003ATLAS Data Challenge
• 200550K SI95, 50 TB disk, data comes by tape
• ?? (ANL,BNL,LBL) distributed organisation.have
  obtained 40 MB/s on network tests
• FNAL(USA CMS) M Kasemann
• US CMS20 CMS
• Raw data and reconstruction at CERN
• 2005.90K SI95, 40 TB disk (and add 40 TB/year)
• All phases of R/D and operation managed and
  operated at FNAL(as for Run 2)
• COSTS
• To 2005 10M ( 35 FTEs)
• Operation 3M /yr. (30 people)

6

• RAL(UK) J Gordon
• Experiments are planning to fund facilities at
  RAL (e.g. CDF)
• Babar have money for 12.5 TB (6 at RAL and rest
  at 9 institutes)
• LHCb (Liverpool) have money for PC MC farm
• RAL await LHC/UK planningexperiments plan
  expenditure as for detectors??
• UK ATLAS want o bid for a regional centre but
  await MONARC results
• Germany A Putzer
• Just beginning national discussions
• Particle Physics would like national centre for
  all 4 experiments (but no money!)
• There will be a proposal from all 4 experiments
  by end 99
• KEK (Japan) K Amako
• Only ATLAS involved (17 institutes)but not all
  convinced of need!
• Human Resources for centre is a big issue
• Currently working on Objectivity evaluations in
  distributed environment
• ITEP Moscow V Kolossov
• Proposal being made for centre at Moscow
  (involving 10 inst. In Moscow area) (but also
  supported by JINR etc.)
• 2106 MIPS (50K SI95), 50 TB disk, 50 PB tape,
  622 Mb/s links
• Prototype this year
• 8 Pentium Pcs (Linux and LHC)

7
Summary points from discussion

• Major issues
• Multiple or single experiment centres??
• Staffing of centres
• How to get funding of all this into experiment
  planning?
• Suggestions for how to move forwards
• Remember it is the experiments who must provide
  the detailed planning and funding
  requests.MONARC is an R/D project providing
  input to experiment planning
• ATLAS/CMS have Technical Reports for computing
  due end 99these should include input from MONARC
  results allowing planning for computing resources
• MONARC workshop May 10-11 will include say 3
  hours on this topic