Improved Energy-Aware Strategies for Periodic Real-Time Tasks under Reliability Constraints

Abstract

This paper revisits the real-time scheduling problem recently introduced by Haque, Aydin and Zhu in a recent issue of this journal [9]. We provide new scheduling strategies that dramatically improve the results of [9]. Résumé : Ce rapport s'intéresseà un problème d'ordonnancement en temps-réelétudié par Haque, Aydin and Zhu et récemment paru dans IEEE TPDS [9]. Nous proposons de nouvelles stratégies de placement et d'ordonnancement qui améliorent significativement les résultats de [9].

Energy-Efficient Partial-Duplication Task Mapping Under Multiple DVFS Schemes

“…or task Replicas (Rep.). A task is executed successfully, if at least one replica is executed without faults [7,15]. The platform has Homogeneous (HO), Heterogeneous (HE) or Single (S) processor(s).…”

“…Heuristics are typically proposed, due to high computation complexity or NP-hard problems, e.g., a first-fit decreasing heuristic for homogeneous platforms [7] and a list scheduling heuristic for heterogeneous platforms [12]. The work in [15] improves the approach proposed in [7] by applying a layered Worst-Fit-Decreasing strategy, where the secondary replicas are executed with same frequency. Hybrid approaches use a given number of replicas to explore the available time slack to impose energy overhead, when an error occurs [25].…”

Development of Energy- Efficient Load Balancing Scheduling Algorithm in Real-time Systems

“…The idea of duplication is to schedule a task graph by mapping some tasks in the set redundantly to reduce the communication between tasks. In [13], the researchers propose an energy efficient real-time scheduling algorithm for reordering chunks in between deadlines and taking advantage of the utilization of the processor to extend deadlines and lower processor frequency hence saving energy. [1] Propose a scheduling technique commensurate with the required improvement.…”