On the analysis of EDF-VD scheduled mixed-criticality real-time systems
Tianyu Zhang, Nan Guan, Qingxu Deng, Wang Yi · 2014
Real-time embedded systems usually integrate multiple functionalities of different criticality levels on a shared hardware platform. For these mixed-criticality real-time systems it is a challenging problem to efficiently utilize system resource to satisfy all the timing constraints on different criticality levels. A simple yet efficient algorithm EDF-VD has recently been proposed to schedule mixed-criticality real-time systems, and shown promising real-time performance. However, the competency of EDF-VD has not been fully exploited due to the imprecise underlying analysis techniques. In this paper, we develop new schedulability analysis methods for EDF-VD. Different from previous analysis methods that separate the analysis on each individual criticality level, our new analysis looks into system behavior crossing multiple criticality levels to obtain more precisely analysis results. Experiments show that our new analysis method can significantly improve guaranteed schedulability of EDF-VD, especially for systems with more criticality levels. The price paid for improved schedulability is higher analysis complexity, but a combination of our new techniques and previous methods can obtain a good balance between the analysis precision and efficiency.