The Particle Physics Data Grid (PPDG) project aims to implement and evaluate distributed (Grid-enabled) data access and management technology for current and future particle and nuclear physics experiments. PPDG has been underway since August 1999, funded initially by the DOE Next Generation Internet program, and recently by a combination of DOE HENP and the Mathematical, Information, and Computational Sciences (MICS) office. It is a collaborative effort between computer scientists and physicists at DOE HENP laboratories (Argonne, Berkeley, Brookhaven, Fermilab, Jefferson and SLAC), Caltech, the San Diego Supercomputer Center (SDSC) and the University of Wisconsin. PPDG has focused on integrating middleware and tools under development by members of the collaboration.
Within the broad vision of Grid-enabled data management and access for HENP, the specific goals of PPDG are to:
- Design, implement, and deploy a Grid-based software infrastructure capable of supporting the data generation, processing and analysis needs common to the physics experiments represented by the participants
- Adapt experiment-specific software to operate in this Grid environment and to exploit this infrastructure. To accomplish these goals, the PPDG will implement and deploy three critical services:
- Reliable and efficient File Replication Service
- High Speed Data Transfer Services
- Multi-Site File Caching and Staging Service
Our accomplishments so far are: deployment and integration of Storage Resource Broker (SRB) catalog and data transfer services with local data access and staging utilities; multithreaded copying of BaBaR data to Lyon; robust file replication rates of up to 80 MBytes/sec between SLAC and Caltech; automated transfer of 1.5 TBytes of Fixed Target data between Fermilab and Indiana. We have collected the short-term file replication needs of the different experiments and have been using them to guide the definition of a File Replication API.
Recently a complementary project named GriPhyN (www.griphyn.org) was proposed to the Information Technology Research (ITR) program of the NSF. Like PPDG, the new project is a joint effort of computer science and physics researchers and spans a large number of universities and research laboratories. If funded, the new project will target an overlapping set of the physics community, concentrate on the long-term IT research necessary for peta-scale data Grids; the GriPhyN proposal hence complements the PPDG effort, which is aimed at providing early experimentation and deployment of Grid services for the HENP community. The early utilization supported by PPDG is serving the purpose of a) providing useful software components, b) educating the physics community on the potential of Grid technology and c) provide feedback to Grid developers to help define the services and architecture.
URLs:
http://www.cacr.caltech.edu/ppdg/
http://gizmo.lbl.gov/ppdg/
http://grid.fnal.gov/ppdg/
http://www.slac.stanford.edu/xorg/ngi/ppdg/ppdg-slac.html
http://lexus.physics.indiana.edu/~focus/PPDG/