Fuzzing Response-Times
Fuzzing-based estimation of WCRT in multi-tasking systems with complex task-interactions.
Primer
Static worst-case response time analysis becomes extremely pessimistic for systems with complex task interactions.
This is because it is not fully context-aware and thus has to consider dynamically infeasible code-paths.
Dynamic testing on the other hand only finds feasible paths, but is not guaranteed to encounter the worst case.
We leverage fuzzing, which is an automated testing technique, to efficiently explore feasible intra- and inter-task control flow paths and produce a close estimate to the WCRT.