Fuzzy based resource management framework for high throughput computing

Abstract

A Fuzzy based scheduling policy was discussed for achieving performance benefits in High Throughput Computing (HTC)-clusters. The scheduling policy was found to consider not only the distributed ownerships and heterogeneous environments but also handles the dynamic state of the cluster. It was found that the algorithm was sensitive to the parameters supplied by the user/ super user which help in achieving high-performance/throughput. The algorithm was suggested to degrade the throughput when the job requirements are not properly provided.