A lazy DVS approach for dynamic real time system

Abstract

This paperpresents a LazyDVS scheduling algorithm that offers higher acceptance ratio (number of tasks able to meet their deadline is to number of tasks appearing) with equivalent orlesser energy consumption for battery powered dynamic real time system. It is modeled for aperiodic workload by utilizing the concept of late start and dynamic voltage scaling. In this paper, with a motivational example we show that existing approaches have poor energy and aperiodic workload management, thus, are not suitable for scheduling aperiodic workload on limited source of energy. Based on this motivation we propose a Lazy DVS scheduling algorithm that tradeoffs between timing constraint and the available energy to provide better acceptance ratio with lower energy. The extensive examples and simulation results illustrate that our approach can effectively improve the acceptance ratio while consuming lower energy.