Compute Resources

The analysis and visualization of large scale data sets requires access to advanced computing resource. Currently, a number of grid computing techniques (see Grid.org projects like Cancer Research, Anthrax Research, Smallpox Research Project, or IBM's Deep Thunder) as well as peer-to-peer (P2P) techniques and applications such as Seti@Home, Einstein@Home, Folding@Home, Astropulse, or Kazaa and hoping for a convergence of the two fields (see recent article by Foster & Iamnitchi) exist.

With support from UITS and SLIS at Indiana University and in collaboration with colleagues from the IR community we are currently creating a Mini-Grid infrastructure for IV and IR research and education. Under consideration are SUN's Grid Engine, TurboWorx, Berkeley Open Infrastructure for Network Computing, LimeWire (BOINC), Condor, Butterfly.net, Twisted and others. This Mini-Grid will be able to scale by harvesting compute cycles from idle (home) computers via a P2P infrastructure, e.g., using XtremWeb. Conforming with Jim Gray's article on Distributed Computing Economics we will aim to put the computation near the data.

Information Visualization Cyberinfrastructure @ SLIS, Indiana University
Last Modified August 2, 2005