A New Fuzzy Algorithm for Global Job Scheduling in Multiclusters and Grids

The computational grid provides a promising platform for the efficient execution of parallel applications. In this paper, we propose a fuzzy algorithm for global job scheduling in multiclusters and grids on the basis of layered task scheduling model. We assume that each job is composed of several parallel tasks, which can be executed concurrently. In addition to CPU requirements of tasks, we consider communication requirements of tasks and network load to find a matching degree between available resources and jobs. Simulation results show the effectiveness of the algorithm in terms of jobs completion time.