Peripheral-conscious energy-efficient scheduling for weakly hard real-time systems

In this paper, we study the problem of reducing the energy consumption for a weakly hard real-time system. The weakly hard real-time system is modelled by the (m, k)-constraints, which require that at least m out of any k consecutive jobs of a task meet their deadlines. Since the system energy is consumed not only by the processor alone but also in a large part by other peripheral devices, we first propose a static approach, with the specifications of peripheral devices taken into consideration, to partition the jobs into mandatory/optional jobs to achieve the dual goals of (m, k)-guarantee and overall energy minimisation. Based on that, we present a dynamic scheduling algorithm that adopts preemption control technique and dynamic mandatory/optional partitioning strategy to reduce the energy consumption of the whole system dynamically. Our approach can effectively minimise the system-wide energy consumption and guarantee the (m, k)-deadlines at the same time. The novelty and effectiveness of our techniques are demonstrated through extensive simulation studies.