Analysis of the effects of weight distribution on proportional-share scheduling algorithms

Abstract

Previously proportional-share scheduling algorithms have been widely studied for guaranteeing the high system reliability. These algorithms allocate CPU resources to tasks based on their weight; however, most proportional-share scheduling algorithms are designed to behave differently according to variations in the overall weight distribution and the number of tasks. In this paper, we classify proportional-share scheduling algorithms according to these differences in behavior and analyze the levels of system reliability they achieve. We examine the effect of the weight on the fairness for different proportional-share scheduling algorithms. Our analysis shows that variations in weight distribution and the number of tasks have a significant impact on the system reliability. Based on this analysis, we further explain which algorithms are best suited to various system environments.