An Improved Adaptive Hierarchical Scheduling Policy for Processor Allocation in Heterogeneous Multicluster Systems
Gursimranjeet Kaur, Amit Chhabra, Harpreet Kaur · 2013
Abstract: Space sharing and time sharing have been traditionally used for scheduling in multiprocessor systems. At medium to heavy loads, time sharing policies perform well. However, at these loads, space sharing policies lead to wastage of resources and premature queuing of jobs, but perform well at low loads. In this paper, we investigate the previous work done in hybrid approaches to schedule in multiprocessor systems which is called hierarchical scheduling. Hierarchical scheduling policies eliminates the short comings of space sharing and time sharing while retaining the advantage of both the policies. We also propose an adaptive hierarchical scheduling policy for scheduling parallel jobs in heterogeneous multicluster systems.