Preemptive Task Partitioning Strategy (PTPS) with Fast Preemption in Heterogeneous Distributed Environment

Efficient preemptions in the scheduling of real time systems cause optimal overhead in parallel computing systems. Periodic and sporadic tasks are exists in the real time systems. The periodic tasks may be divided into the synchronous and asynchronous categories. The management of the resource sharing in the parallel computing can be powerfully achieved by preemptive scheduling. Fast preemptions are necessary to achieve the high degree parallelism. In this paper, Earliest Starting Time parameter expended up to a large degree of heterogeneity. We compare the proposed algorithm with the existed well known algorithms: preemptive MCP and FPS algorithms. The result shows better performance of the PTPS in terms of average NSL and running time complexities.