Fault-aware grid scheduling using performance prediction by workload modeling
The Journal of Supercomputing
Hi-index | 0.02 |
In high throughput computing environments, anopportunistic scheduling policy is used for placement ofbatch jobs on idle workstations for execution. In thispaper, we propose a new opportunistic scheduling policywith the aim of exploiting the idle resources provided bysystems such as Condor in an opportunistic manner. Wecompared the performance of the proposed schedulingpolicy with three policies namely the scheduling policyused in Condor, job rotate and round robin policiesthrough simulation. The results show that the proposedscheduling policy offers substantial performanceimprovements over the exiting approaches. Furthermore,this improved performance is achieved without loss ofthroughput and it results in a more interactive nature ofthe system thus increasing its appeal.