Analysis of the effects of weight distribution on proportional share scheduling

Proportional share scheduling algorithms have been widely studied as means of guaranteeing the QoS of applications. These algorithms allocate CPU resources to tasks based on their weight, but most are designed to behave differently under variations in overall weight distribution and number of tasks. In this paper, we classify proportional share scheduling algorithms according to these differences in behavior and analyze the levels of QoS they achieve. Our analysis shows that variations in weight distribution and number of tasks have a significant impact on QoS. Based on this analysis, we further explain which algorithms are best suited to various environments.

[1]  Hussein M. Abdel-Wahab,et al.  A proportional share resource allocation algorithm for real-time, time-shared systems , 1996, 17th IEEE Real-Time Systems Symposium.

[2]  Carl A. Waldspurger,et al.  Lottery and stride scheduling: flexible proportional-share resource management , 1995 .

[3]  Jason Nieh,et al.  Group Ratio Round-Robin: O(1) Proportional Share Scheduling for Uniprocessor and Multiprocessor Systems , 2005, USENIX Annual Technical Conference, General Track.

[4]  Hui Zhang,et al.  WF/sup 2/Q: worst-case fair weighted fair queueing , 1996, Proceedings of IEEE INFOCOM '96. Conference on Computer Communications.

[5]  Jason Nieh,et al.  Virtual-Time Round-Robin: An O(1) Proportional Share Scheduler , 2001, USENIX Annual Technical Conference, General Track.

[6]  George Varghese,et al.  Efficient fair queueing using deficit round-robin , 1996, TNET.