Energy reduction in weakly hard real time systems

Abstract

Real Time Systems can be classified into Hard, Soft and Firm Real Time Systems depending on the consequences of a task missing its deadline. A Hard Real Time System produce required results before the specified time bound. In Soft Real Time Systems the few misses of deadline is acceptable if they have no harm. The Generalized case of Real Time Systems is Weakly Hard Real Time System which is motivation of the observation for real time applications when some deadline misses are acceptable as long as they are spaced evenly. This paper is an attempt to provide a state-of-the-art review of basic model of Real Time System, clock driven scheduling, priority driven scheduling and proposes an Inverse Rate Monotonic algorithm. The Inverse Rate Monotonic algorithm effectively reduce the energy by 15% along with Dynamic Voltage Scaling strategy which provides significant energy savings while maintain real time deadline guarantee. A tool has been created in C language to observe the scheduling and energy consumption of task by Inverse Rate Monotonic.