WeaklyHard.jl: Scalable Analysis of Weakly-Hard Constraints

Weakly-hard models have been used to analyse real-time systems subject to patterns of deadline hits and misses. However, the tools that are available in the literature have a set of shortcomings. The analysis they offer is limited to a single weaklyhard constraint and to patterns that specify the number of misses, rather than the number of hits. Furthermore, the scalability of the tools is limited, effectively making it hard to address systems where deadline misses are really sporadic events. In this paper we present WeaklyHard.jl, a scalable tool to analyse a set of weakly hard constraints belonging to all the four types of weakly hard models. To achieve scalability, we exploit novel dominance relations between weakly-hard constraints, based on deadline hits. We provide experimental evidence of the tool’s scalability, compared to the state-of-the-art for a single constraint, a thorough investigation of hit-based weakly-hard constraints, and a sensitivity analysis to constraint set parameters.