Research Infrastructure: The Scalable Intracampus Research Grid for Computer Science Research

University Of Tennessee Knoxville, Knoxville TN

Investigators

Abstract

EIA-9972889 Jack J. Dongarra Micah D. Beck Michael W. Berry Jens Gregor Michael A. Langston University of Tennessee CISE Research Infrastructure: The Scalable Intracampus Research Grid for Computer Science Research The University of Tennessee (Knoxville) will build a computational grid at the University that will support their research in grid computing systems as well as advance applications that the grid will eventually support. The campus grid will consist of several grid service clusters - cycle servers linked in an advanced network. Grid research will address middleware to support performance based management that can seek out most appropriate software and other services, support distributed data management, process migration, load balancing, scheduling and fault-tolerance. Middleware will be based on their existing NetSolve project. Other systems supporting the grid include the Internet Backplane Protocol for fine-grained naming and distributed storage, checkpointing research, and the Network Weather Service for performance monitoring. Applications to be supported with the developing grid include medical imaging, materials design (such as lubricants), computational ecology, and machine design. Applications will benefit from the full capabilities of several clusters as well as provide challenging applications to push the grid programming interface and technology.