AN APPROACH TO PROCESSOR ALLOCATION SCHEDULING FOR HETEROGENEOUS MULTI-CLUSTER SYSTEM

The requirement of high computational tasks which requires large job sizes has raised, due to this the demand of supercomputer increases which could handle large computational requirements. The fact-finding is made on evolving an Adaptive Scheduling Policy for Non-dedicated Heterogeneous multi-cluster system. This adaptive scheduling policy which will raises the performance of multi-cluster compared to existing policies Multi-cluster are systems are used now a days to get high performance and higher system utilizations .Two major factors for effective scheduling namely, Job scheduling and Processor allocation have been discussed and accordingly the algorithm is designed. The basis of the main issues of algorithm in heterogeneous clusters such as Resource Fragmentation and Speed Heterogeneity. The scheduling policies are there to allocate the processor to each job and adaptive scheduling is one of the scheduling which deals with these jobs that require suitable processor. In this research is base on Adaptive Scheduling Policy for Heterogeneous Multi-cluster system. Furthermore, the performance of the proposed algorithm approach dealing with Multi-site execution is compared with the case of Multi-pool configuration in which different clusters in the computing grid are viewed as different processor pools and each job must be allocated to a set of processors belonged to exactly one of these pools. No jobs can continuously use processors from various pools. So, if the first job in queue could not fit into any single site in the grid, it would have to wait.