Energy efficient EDF-VD-based mixed-criticality scheduling with shared resources

Abstract

In this paper, we consider simultaneously energy consumption and resource synchronization in mixed-criticality (MC) single processor systems. First, we give a feasibility analysis of single processor systems that execute real-time MC tasks. Second, an energy efficient speed based on sufficient feasibility condition is computed. In addition, we propose a single speed energy efficient algorithm for MC tasks with shared resources (ASS). The ASS algorithm scheduling tasks with SLO is too conservative with much room to save more energy. For energy efficiency, dual speeds scheduling algorithm (DSS) based on ASS algorithm is proposed. It schedules tasks at SLO with blocking and SLO′ without blocking. Blocking means that the higher priority tasks cannot preempt the execution of current low-priority tasks because low-priority tasks occupy the shared resources required for the higher priority tasks. Moreover, we prove that the DSS algorithm is feasible. Finally, the real-life synthetic application and extensive simulation are applied to validate the proposed algorithm. The experimental results show that the DSS algorithm can reduce energy consumption up to 11.82% compared with the existing approaches.