GriPhyN

1
GriPhyN Towards Petascale Virtual Data
GridsPaul Avery, University of Florida Ian
Foster, University of ChicagoNSF Grant
ITR-0086044 www.griphyn.org

• Research Objectives
• Provide tools to manage, discover, apply
  peta-scale data and software resources
• Enable Grids to be used by scientists as simply
  as personal workstations
• Approach
• A uniform virtual data model (VDL) to express
  datasets and workflow. Advanced planners,
  executors, monitors, predictors, and fault
  recovery to make the Grid like a workstation
• Intensive Experiment/CS collaboration
• Partnership with iVDGL and PPDG to deploy
  international Grid www.ivdgl.org/grid2003
• Broader Impact
• Unified language to express data transformation
  benefits all sciences
• Lowered barrier to computation increased
  collaboration and communication
• Benefits to education enhance discovery and
  learning of scientific processes
• Participants
• Argonne, Boston U, Brookhaven, Caltech,
  Fermilab,Hampton, Johns Hopkins, Harvard, Salish
  Kootenai, Texas AM, UC Berkeley, UC San
  Diego,U Chicago, U Florida, USC/ISI, U Texas, U
  Wisconsin

• Significant Results
• Virtual Data Language, Virtual Data Toolkit (VDT)
• VDT tools applied by multiple sciences (ATLAS,
  CMS, LIGO, SDSS, Bioinformatics, others )
• Collaboration with other NSF DOE supported
  projects to build international Grid Grid2003
• Outreach impact Virtual data for
  QuarkNet,Summer Grid Workshop 2004-05
• Intensive student involvement 25 graduate, 2
  undergraduate 4 CS PhDs as of 07/2005