Summary: The Case for Fair Multiprocessor Scheduling #
Anand Srinivasan, Philip Holman, James H. Anderson, and Sanjoy Baruah
Department of Computer Science
University of North Carolina at Chapel Hill
E­mail: {anands,holman,anderson,baruah}@cs.unc.edu
November 2002
Abstract
Partitioning and global scheduling are two approaches for scheduling real­time tasks on multi­
processors. Though partitioning is sub­optimal, it has traditionally been preferred; this is mainly
due to the fact that well­understood uniprocessor scheduling algorithms can be used on each proces­
sor. In recent years, global­scheduling algorithms based on the concept of ``proportionate fairness''
(Pfairness) have received considerable attention. Pfair algorithms are of interest because they are
currently the only known method for optimally scheduling periodic, sporadic, and ``rate­based'' task
systems on multiprocessors. In addition, there has been growing practical interest in scheduling with
fairness guarantees. However, the frequency of context switching and migration in Pfair­scheduled
systems has led to some questions concerning the practicality of Pfair scheduling.
In this paper, we investigate this issue by comparing the PD 2 Pfair algorithm to the EDF­FF
partitioning scheme, which uses ``first fit'' (FF) as a partitioning heuristic and the earliest­deadline­
first (EDF) algorithm for per­processor scheduling. We present experimental results that show that
PD 2 is competitive with, and in some cases outperforms, EDF­FF. These results suggest that Pfair

Source: Anderson, James - Department of Computer Science, University of North Carolina at Chapel Hill

Collections: Computer Technologies and Information Sciences