Task period selection and schedulability in real-time systems

Abstract

In many real time applications, especially those involving computer controlled systems, the application tasks often have a maximal acceptable latency, and small latency is preferred to large. The interaction between choosing task periods to meet the individual latency requirements and scheduling the resulting task set was investigated by D. Seto et al. (1996) using dynamic priority scheduling methods. We present algorithms based on static priority scheduling methods to determine optimal periods for each task in the task set. The solution to the period selection problem optimizes a system wide performance measure, subject to meeting the maximal acceptable latency requirements of each task. The paper also contributes to a new aspect of rate monotonic scheduling, the optimal design of task periods in connection with application related timing specifications and task set schedulability.