Semi-clairvoyant scheduling in non-preemptive fixed-priority mixed-criticality systems

Abstract

Non-preemptive real-time scheduling of mixed-criticality systems (MCSs) in which tasks have different levels of criticality has gained increasing attention. In this paper, we propose schedulability tests specifically designed for the imprecise mixed-criticality task model, which is the first effort for non-preemptive fixed-priority semi-clairvoyant scheduling (NPFP-SC). It schedules tasks by a non-preemptive fixed-priority scheme while the system transition becomes apparent upon the arrival of high-criticality tasks. We next propose the energy-aware scheduling algorithm based on NPFP-SC schedulability tests, called EA-NPFP-SC, to solve the energy problem of MCSs. Our experimental results indicate that the NPFP-SC algorithm outperforms existing methods in terms of schedulability ratio by 17.33%, while the EA-NPFP-SC algorithm achieves a 40.18% reduction in energy consumption compared to the NPFP-SC algorithm.