Summary: Fair Scheduling of Dynamic Task Systems on
Anand Srinivsan and James H. Anderson
Department of Computer Science, University of North Carolina
Chapel Hill, NC 275993175 USA
In dynamic realtime task systems, tasks that are subject to deadlines are al
lowed to join and leave the system. In previous work, Stoica et al. and Baruah et al.
presented conditions under which such joins and leaves may occur in fairscheduled
uniprocessor systems without causing missed deadlines. In this paper, we extend
their work by considering fairscheduled multiprocessors. We show that their con
ditions are su#cient on M processors, under any deadlinebased Pfair scheduling
algorithm, if the utilization of every subset of M - 1 tasks is at most one. Further,
for the general case in which task utilizations are not restricted in this way, we
derive su#cient join/leave conditions for the PD 2 Pfair algorithm. We also show
that, in general, these conditions cannot be improved upon without causing missed
Key words: Dynamic task systems, Pfairness, multiprocessor, realtime scheduling
In many realtime systems, the set of runnable tasks may change dynamically.