Rate-monotonic scheduling for reducing system-wide energy consumption for hard real-time systems

Abstract

In this paper, we present system-wide dynamic scheduling algorithms to reduce the energy consumption by both the core DVS processor and multiple non-DVS peripheral devices for hard real-time systems scheduled with rate-monotonic scheduling (RMS) scheme. In our research, we first present an approach to leverage the use of the critical speed strategy and the traditional DVS strategy based on the job workload to be finished within certain interval. Then dynamic scheduling approaches are proposed in the management of speed determination and device shut-down to reduce the energy at the system level. Compared with existing research, our approach can effectively reduce the overall energy consumption for both CPU and peripheral devices.