Title: ANALYSIS OF FIXED PRIORITY SWAP SCHEDULING POLICY FOR REAL-TIME AND NON REAL-TIME JOBS

Issue Number: Vol. 2, No. 3
Year of Publication: Jul - 2012
Page Numbers: 488-495
Authors: Michael Okopa, Tonny Bulega
Journal Name: International Journal of New Computer Architectures and their Applications (IJNCAA)
- Hong Kong

Abstract:


Scheduling of real-time jobs is challenging given the fact that they are delay sensitive. In order to improve the performance of such jobs, a number of scheduling policies have been proposed. SWAP, a non-preemptive threshold based policy has been suggested to approximate the Shortest Job First (SJF) policy which is a non-preemptive optimal policy. SWAP differentiates between short and large jobs such that short jobs are given precedence over large jobs. However, SWAP cannot guarantee the service quality of real-time jobs which are constrained in terms of deadlines especially under heavy load or congestion. In this paper, we analyze a variant of SWAP called Fixed Priority SWAP policy that can provide service differentiation so as to reduce the delay experienced by real-time jobs without hurting much the performance of non real-time jobs. The service differentiation attribute is based on the time sensitivity of the jobs, where more time sensitive jobs also called real-time jobs are given higher priority over less time sensitive jobs also termed non real-time jobs. We evaluate Fixed Priority SWAP under an M/G/1 queue for job size distributions with varying coefficient of variation (CoV) to analyze the impact of service differentiation on real-time and non real-time jobs. Our detailed numerical analysis reveals that Fixed Priority SWAP policy improves the performance of real-time jobs at the expense of little or no penalty for the non real-time jobs based on the threshold value.