Scheduling Algorithms for Elastic Mixed-Criticality Tasks in Multicore Systems (Extended Version)
The Elastic Mixed-Criticality (E-MC) task model and an Early-Release EDF (ER-EDF) scheduling algorithm have been studied to address the service interruption problem for low-criticality tasks in uniprocessor systems, where the minimum service requirements of low-criticality tasks are guaranteed by their maximum periods. In this paper, focusing on multicore systems, we first investigate and empirically evaluate the schedulability of E-MC tasks under partitioned-EDF (P-EDF) by considering various task-to-core mapping heuristics. Then, to improve the service levels of low-criticality tasks by exploiting slack at runtime, with and without task migrations being considered, we study both global and local early-release schemes. Moreover, considering varied migration overheads between different cores, we propose the multicore-aware and migration constrained global-ER schemes. The simulation results show that, compared to the state-of-the-art Global EDF-VD scheduler, P-EDF with various partitioning heuristics can lead to many more schedulable E-MC task sets. Moreover, our proposed global and local ER schemes can significantly improve the execution frequencies (and thus service levels) of low-criticality tasks, while Global EDF-VD may severely affect them by discarding most of their task instances at runtime (especially for systems with more cores). Furthermore, by allowing task migrations, global-ER schemes can dramatically improve low-criticality tasks’ service levels for partitionings that do not balance high- and low-criticality tasks among the cores.