A dynamic job scheduling algorithm for computational grid

Abstract

In this paper, a dynamic job-scheduling algorithm is proposed for a computational grid of autonomous nodes. This algorithm tries to utilize the information of a practical system to allocate the jobs more evenly. In this algorithm, the communication time between nodes and the scheduler is overlapped with the computation time of the nodes. So the communication overhead can be little. The principle of scheduling the job is based on the desirability of each node. The scheduler would not allocate a new job to a node that is already fully utilized. The execution efficiency of system will be increased. An implementation framework of the algorithm is also introduced. © Springer-Verlag 2004.