
Virtual Grid Application Development Software
Overview
Grids hold great promise for researchers who need to connect to remote computers, databases and other resources, however, they can be complex, unreliable, and require a huge time investment for even low-level operations. To keep these problems from impeding the potential of grids and distributed resources, the Virtual Grid Application Development Software (VGrADS) project was developed. VGrADS attacks a fundamental part of the grid usability problem: how to more effectively program these highly complex and dynamic systems. VGrADS is developing software tools that simplify and accelerate the development of grid applications and services, while delivering high levels of performance and resource efficiency. VGrADS is based on the earlier GrADS project.
The RENCI Contribution
As part of the VGrADS effort, RENCI looks at temporal-based reasoning to qualitatively assess, diagnose and adapt long-running applications. RENCI also investigates qualitative metrics for implementing a multi-level fault tolerance strategy, especially in the context of workflows that have strict deadlines, such as weather forecasting. RENCI is also involved in implementing and evaluating different fault-tolerance and recovery mechanisms for such workflows.
Funding
Cooperative Agreement issued to Rice University under National Science Foundation Cooperative Agreement No. CCR-0331645 with a sub agreement to the University of North Carolina at Chapel Hill.
Daniel A. Reed, Wei Deng, Celso Mendes, et al. "New Grid Scheduling and Rescheduling Methods in the GrADS Project," Proceedings of the International Parallel and Distributed Processing Symposium Workshop (IPDPS NGS) IEEE Computer Society Press, 2003.
Presentations
Daniel A. Reed, presentation on Fault Tolerance at VGrADS All Hands Meeting, September 2005
Daniel A. Reed, presentation at the VGrADS Site Visit, April 2005
Partners
Rice University
University of California at San Diego
University of California at Santa Barbara
University of Houston
University of North Carolina at Chapel Hill
University of Southern California Information Sciences Institute
University of Tennessee at Knoxville
RENCI
Links
VGrADS Project Website