Abstract

This paper deals with dynamic scheduling in realtimesystems that have Quality of Service requirements.We assume that tasks are periodic and may miss theirdeadlines, occasionally, as defined by the so-called Skip-Over model. In this paper, we present a dynamic schedulingalgorithm, called RLP (Red as Late as possible, a variant ofEarliest Deadline to make slack stealing and to get betterperformance in terms of ratio of periodic task instanceswhich complete before their deadline). Simulation resultsshow that RLP outperforms the two conventional skip-overalgorithms, namely RTO and BWP, introduced about tenyears ago. Then, we investigate a second criteria calledfairness, aiming to balance individual success ratios of tasks.We present variants of RLP to improve fairness and reportcomparative simulation results. Finally, we present the integrationof these QoS scheduling services into CLEOPATRE1,a free open-source library which offers selectable real-timefacilities on shelves.