Energy-aware reliability guarantee scheduling with semi-clairvoyant in mixed-criticality systems

The management of energy in mixed-criticality systems (MCS) has been widely accomplished through Dynamic Voltage and Frequency Scaling (DVFS) techniques. Nevertheless, recent studies indicated that the DVFS has negative impact on the reliability of the MCS. In this work, we investigate the problem of reliability-aware power management (RAPM) for semi-clairvoyant MCS with the objective of saving energy while meeting both reliability and deadline constraints. We first address the RAPM problem in semi-clairvoyant MCS with the imprecise mixed-criticality task model. Then, we analyze the feasibility issue of MCS under the constraints deadline and reliability using the Demand Bound Function and derive sufficient conditions of the schedulability test. Based on the analysis, we propose an energy-aware reliability guarantee scheduling algorithm, called EARGS, which reduces energy consumption while satisfying both the deadline and reliability constraints. Finally, the experiment results indicate that the EARGS algorithm saves approximately 25.80 % of energy consumption compared to other state-of-the-art methods.