Reliability guaranteed energy-aware frame-based task set execution strategy for hard real-time systems

摘要

In this paper, we study the problem of how to execute a real-time frame-based task set on DVFS-enabled processors so that the system's reliability can be guaranteed and the energy consumption of executing the task set is minimized. To ensure the reliability requirement, processing resources are reserved to re-execute tasks when transient faults occur. However, different from commonly used approaches that the reserved processing resources are shared by all tasks in the task set, we judiciously select a subset of tasks to share these reserved resources for recovery purposes. In addition, we formally prove that for a give task set, the system achieves the highest reliability and consumes the least amount of energy when the task set is executed with a uniform frequency (or neighboring frequencies if the desired frequency is not available). Based on the theoretic conclusion, rather than heuristically searching for appropriate execution frequency for each individual task as used in many research work, we directly calculate the optimal execution frequency for the task set. Our simulation results have shown that with our approach, we can not only guarantee the same level of system reliability, but also have up to 15% energy saving improvement over other fault recovery-based approaches existed in the literature. Furthermore, as our approach does not require frequent frequency changes, it works particularly effective on processors where frequency switching overhead is large and not negligible.