An energy-aware scheduling for real-time task synchronization using DVS and leakage-aware methods

Abstract

Due to the importance of resource allocation and energy efficiency, this paper considers minimizing priority inversion and energy consumptions in the embedded real-time systems. While dynamic voltage scaling (DVS) is known to reduce dynamic power consumption, it also causes increased blocking time of lower priority tasks and leakage energy consumption due to increased execution. We proposed a concept of latency locking to prevent priority inversion using sleeping mode and define a block-free interval in which both DVS and leakage-aware methods can be applied. In order to compute the optimal sleeping time and its duration and to meet the timing constraints, we also propose a weighted directed graph (WDG) to obtain additional task information. By traversing WDG, task information can be updated online and the scheduling decisions could be done in linear time complexity.