An Improved Job Scheduling Algorithm Based on the First Distribution and Redistribution Strategy for Grid Computing

Abstract

As the traditional job scheduling algorithms ignore the impact of resource fragmentation, there are some insufficiencies in terms of the job execution time for grid computing. Since deficiency occurs to these traditional job scheduling algorithms , why not reduce the resource fragmentation probability by the way of combining backfilling with the priority scheduling strategy? If it should be, how to be regulated? This paper argues that the treatment of resource recovery should be the solution to a considerable resource fragmentation problem. We present an improved job scheduling algorithm which optimizes the traditional grid scheduling algorithms based on the first distribution and the redistribution strategy of jobs. The algorithm considers the resource recycling so specifically that it can distribute the job efficiently. Simulation experiment results demonstrate that the improved job scheduling algorithm can greatly improve the resource utilization rate; at meanwhile, it can also track the system throughput rate with excellent performance.