A Blocking-Aware Scheduling for Real-Time Task Synchronization Using a Leakage-Controlled Method

Abstract

Due to the importance of power dissipation in the wireless sensor networks and embedded systems, real-time scheduling has been studied in terms of various optimization problems. Real-time tasks that synchronize to enforce mutually exclusive access to the shared resources could be blocked by lower priority tasks. While dynamic voltage scaling (DVS) is known to reduce dynamic power consumption, it causes increased blocking time due to lower priority tasks that prolong the interval over which a computation is carried out. Additionally, processor slowdown to increase execution time implies greater leakage energy consumption. In this paper, a leakage-controlled method is proposed, which decreases both priority inversion and power consumption. Based on priority ceiling protocol (PCP) and a graph reduction technique, this method can decrease more energy consumption and avoid priority inversion for real-time tasks.