Fixed-priority global scheduling for mixed-criticality real-time systems

There has been a growing interest in recent years in mixed-criticality real-time systems in which tasks are attributed different levels of criticality based on the degree to which their deadlines must be assured. While most of the initial efforts on mixed-criticality systems targeted single-processor systems, the research community has recently started to investigate multiprocessor mixed-criticality real-time systems. In this paper we investigate how global, fixed-priority algorithms can be applied in the context of multiprocessor mixed-criticality systems. We identify the two key dimensions of the problem - priority assignment and schedulability testing, and consider candidate algorithms for each dimension. We also propose a new and simple priority assignment policy called CPRatio that considers both task criticality and timing constraints to improve the schedulability of mixed-criticality task sets. We experimentally evaluate the performance of priority assignment strategies and schedulability tests in the global multiprocessor mixed-criticality scheduling context.