The heterogeneity and unpredictability of grid computing environments is a severe challenge to algorithm selection for large-scale scientific codes. This paper discusses a set of requirements for an algorithm selection web service, which supports runtime, context-aware algorithm recommendations in a multi-language, multi-platform environment. An architecture and prototype implementation of such a system is described, along with a perspective of how such a system will be helpful to other grid middleware such as meta-scheduling systems.
Citation:
Prachi Bora, Calvin Ribbens, Sandeep Prabhakar, Gautam Swaminathan, Malarvizhi Chinnusamy, Ashwin Jeyakumar, Beatriz Diaz-Acosta, "Issues in Runtime Algorithm Selection for Grid Environments," clade, pp.80, International Workshop on Challenges of Large Applications in Distributed Environments, 2003