Job scheduling in a grid cluster

This paper studies the performance of a distributed system. In particular, our work focuses on the study and comparison of job scheduling techniques in a cluster that could be part of a computational Grid. We examine two different job allocation policies, one static and one dynamic, and three job scheduling policies combined together. The performance of different scheduling schemes is compared over various workloads. The purpose of this study is to achieve the maximum efficiency of the system through the discovery of an efficient operating scenario. Simulation results reveal that the scheduling scheme that should be chosen depends on whether load information is available and also whether in advance knowledge of jobs service time is available.