Enhanced Fault-Tolerant Fixed-Priority Scheduling of Hard Real-Time Tasks on Multi-core Platforms

In this paper, we study the problem of partitioned scheduling of periodic real-time tasks with the capability of tolerating transient faults on multi-core platforms under Rate Monotonic Scheduling (RMS) policy. In our approach, we exploit the implicit relations among periods and recovery costs among tasks and develop a novel metric, called "compatibility index", to quantify how "compatible" a task set is when they are allocated on the same core. We theoretically analyze its properties for improving the system schedulability. Based on this metric, we propose two task partitioning schemes to partition hard real-time tasks with fault-tolerance requirements on multi-core platforms. Simulation results demonstrate that our proposed approaches can significantly enhance the performance of existing techniques.