A statistical response-time analysis of complex real-time embedded systems by using timing traces

Real-time embedded systems are becoming ever more complex, and we are reaching the stage where even if static Response-Time Analysis (RTA) was feasible from a cost and technical perspective, the results are overly pessimistic making them less useful to the practitioner. When combined with the fact that most timing analysis tends to be statistical in nature, this suggests there should be a move toward statistical RTA. However, to make such analysis useful, it is imperative that we have evidence that the statistical RTA and the information analyzed is sufficiently accurate. In this paper we present and validate a technique for statistical RTA that can cope with systems that are complex from both a size and tasks' dependencies perspective. This claim is backed up by our evaluation using information from real industrial control systems.