Importance Sampling of Runtime Interference

Behavior of open systems is subject to dynamically changing external environment. Runtime monitoring and checking methods are employed to detect interference of functional features in these open systems. Such checking usually has a large impact on the runtime performance. The overhead of checking is unavoidable for fatal anomalies. For less severe anomalies, the performance penalty is not acceptable if the occurrence rate is very small. This paper proposes a new method of estimating the occurrence rate of rare events at runtime with the importance sampling. It is the first report of applying the method to running systems, instead of using discrete simulation models or design models.