Real-Time Scheduling on Uni- and Multiprocessors based on Priority Promotions

This paper addresses the problem of real-time scheduling of a set of sporadic tasks on uni- and multiprocessor platform based on priority promotion. A new preemptive scheduling algorithm, called Fixed-Priority with Priority Promotion (FPP), is proposed. In FPP scheduling, tasks are executed similar to traditional fixed-priority (FP) scheduling but the priority of some tasks are promoted at fixed time interval (called, promotion point) relative to the release time of each job. A policy called Increase Priority at Deadline Difference (IPDD) to compute the promotion points and promoted priorities for each task is proposed. FPP scheduling prioritizes jobs according to Earliest-Deadline-First (EDF) priority when all tasks' priorities follow IPDD policy.It is known that managing (i.e., inserting and removing) jobs in the ready queue of traditional EDF scheduler is more complex than that of FP scheduler. To avoid such problem in FPP scheduling, a simple data structure and efficient operations to manage jobs in the ready queue are proposed. In addition, techniques for implementing priority promotions with and without the use of a hardware timer are proposed.Finally, an effective scheme to reduce the average number of priority promotions is proposed: if a task set is not schedulable using traditional FP scheduling, then promotion points are assigned only to those tasks that need them to meet the deadlines; otherwise, tasks are assigned traditional fixed priorities without any priority promotion. Empirical investigation shows the effectiveness of the proposed scheme in reducing overhead on uniprocessor and in accepting larger number of task sets in comparison to that of using state-of-the-art global schedulability tests for multiprocessors.