An energy-aware algorithm exploiting limited preemptive scheduling under fixed priorities

Abstract

This paper presents a new energy-aware algorithm that integrates Dynamic Voltage and Frequency Scaling (DVFS) and Dynamic Power Management (DPM) techniques to further reduce energy consumption in embedded systems. It consists of an off-line DVFS-stage, for computing the speed that minimizes energy consumption during active intervals while guaranteeing timing constraints, and an online DPM-stage, for prolonging sleep intervals by postponing task execution. Moreover, limited preemptive scheduling is exploited to reduce preemption costs and further extend sleep intervals under fixed-priority systems, with respect to fully preemptive schedulers. The online algorithm has a constant complexity and preemption costs are taken into account in the analysis. A set of simulation experiments are reported to illustrate the behavior of the proposed approach as a function of different parameters and compare its performance with the state-of-art methods available in the literature.