Energy-efficient partitioned semi-clairvoyant scheduling in mixed-criticality system with graceful degradation
摘要
Previous works have focused on non-clairvoyant task models in mixed-criticality systems (MCS) with multiprocessor platforms, where the completion of high-criticality (HI) tasks determines the behavioral changes of the system. In this research, we explore semi-clairvoyant scheduling in MCS with graceful degradation. Jobs will have the ability to know whether their execution time will exceed the worst-case execution time in the low-criticality (LO) mode at their arrival time. We propose an energy-aware algorithm to determine the optimal speed for energy efficiency in the LO mode on each processor, and then we present a novel energy-efficient partitioned semi-clairvoyant scheduling algorithm, named EEPSCMC, aimed at reducing the energy consumption of the system. We perform experiments to evaluate EEPSCMC in comparison with four other heuristic algorithms, and the experimental results demonstrate that EEPSCMC outperforms other algorithms, exhibiting superior performance in terms of normalized energy consumption, saving up to