An Efficient Load-Sharing and Fault-Tolerance Algorithm in Internet-Based Clustering Systems

摘要

This paper proposes an efficient algorithm for load-sharing and fault-tolerance in Internet-based clustering systems. The algorithm creates a global scheduler based on the Weighted Factoring algorithm. And it applies an adaptive granularity strategy and the refined fixed granularity algorithm for better performance. It may also execute a partial job several times for fault-tolerance. For the simulation, the matrix multiplication using PVM is used in a Internet-based clustering system. Compared to other algorithms such as Send, GSS and Weighted Factoring, the proposed algorithm results in an improvement of performance by 55%, 63% and 20%, respectively. Also, this paper shows that it can process the fault-tolerance.