Dynamic task scheduling algorithm with task synchronization and energy savings in hard real-time systems

Abstract

HDSA（hybrid dynamic scheduling algorithm） was proposeed,which slows down processor via voltage/frequency scaling,such that the real-time schedulability and task synchronization can be satisfied concurrently with as minimum as possible energy consumption.Based on RM and EDF,HDSA consists of a static algorithm and a dynamic one.The static algorithm computes the initially static speeds of all tasks under the static conditions.The dynamic algorithm fixes the execution frequencies of all critical sections at runtime,and reclaims the slack of tasks during their executions so as to further reduce energy consumption while satisfying timing constraints.On the other hand,frequent voltage switching,which is caused by the frequency inheritance,can be avoided by the approach when a task with a lower priority blocks a task with a higher priority.Consequently,the scheduling costs can also be reduced efficiently.The experimental results illustrate that the proposed algorithm obviously outperforms existing algorithms in terms of energy consumption.

Keywords

• computer application and technology;real-time systems;dynamic scheduling algorithm;task synchronization;energy savings