PPA hopes to get ~$600k per year to rebuild and grow the computing capacity developed for BABAR. The idea is to serve projects/groups without their own hardware purchases and to act as a reservoir for cycles for one-off tasks (eg a data reprocessing needing a peak of cycles).

Here are initial issues to be discussed:

  • determining static allocations - and who are the players?
    • EXO, CDMS, LBNE, HEP Theory, new DM, LSST, SiD, DES; KIPAC users
    • some BABAR allocation during ramp. LTDA is their home
    • KIPAC Theory; Steve Allen et al 
    • Fermi users (on Fermi nickel?) Anyone else?
  • mechanism to borrow allocation
  • determining annual needs (e.g. CPU vs disk)
  • encouraging participation from projects (incentives to join?)
  • point person + advisory (SCA Advisory?)
  • usage reporting
  • No labels