Integration of safety-critical tasks with different certification requirements onto a common hardware platform has become a growing tendency in the design of real-time and embedded systems. In the past decade, great efforts have been made to develop techniques for handling uncertainties in task worst-case execution time, quality-of-service, and schedulability of mixed-criticality systems. However, few works take fault-tolerance as a design requirement. In this paper, we address the scheduling of fault-tolerant mixed-criticality systems to ensure the safety of tasks at different levels of criticalities in the presence of transient faults. We adopt task re-execution as the fault-tolerant technique. Extensive simulations were performed to validate the effectiveness of our algorithm. Simulation results show that our algorithm results in up to [Formula: see text] and [Formula: see text] improvement in system reliability and schedule feasibility as compared to existing techniques, which contributes to a more safe system.
