A Three Phase Scheduling for System Energy Minimization of Weakly Hard Real Time Systems
This paper aims to present a three phase scheduling algorithm that offers lesser energy consumption for weakly hard real time systems modeled with (𝕞𝕞𝕞𝕞, 𝕜𝕜𝕜𝕜) constraint. The weakly hard real time system consists of a DVS processor (frequency dependent) and peripheral devices (frequency independent) components. The energy minimization is done in three phase taking into account the preemption overhead. The first phase partitions the jobs into mandatory and optional while assigning processor speed ensuring the feasibility of the task set. The second phase proposes a greedy based preemption control technique which reduces the energy consumption due to preemption. While the third phase refines the feasible schedule received from the second phase by two methods, namely speed adjustment and delayed start. The proposed speed adjustment assigns optimal speed to each job whereas fragmented idle slots are accumulated to provide better opportunity to switch the component into sleep state by delayed start strategy as a result leads to energy saving. The simulation results and examples illustrate that our approach can effectively reduce the overall system energy consumption (especially for systems with higher utilizations) while guaranteeing the (𝕞𝕞𝕞𝕞, 𝕜𝕜𝕜𝕜) at the same time.
