An enhanced grid scheduling with job priority and equitable interval job distribution

The scheduling service is an important component of large scale computing environments. In this paper, we take a local and grid-wise look at the scheduling problem. First an advance backfilling algorithm based on the job square with a wide job priority is presented. Experimental results show that the priority scheduler reduces the mean waiting time to an extent that depends on the proportion of narrow jobs within a workload. Subsequently, we consider a load sharing technique that selects the site in the Grid that is executing the least number of jobs of similar size as that of the current job. The adaptive sharing scheme offers significant benefits in terms of the average weighted waiting time.