Towards achieving efficient runtime enforcement for component based systems

Abstract

Formally ensuring the correctness of component-based, concurrent systems is an arduous task, mainly because exhaustive methods such as model-checking quickly run into state-explosion problems; this is typically caused by the multiple thread interleavings of the system being analysed, and the range of data the system can input and react to. Runtime Verification (RV) is an appealing compromise towards ensuring correctness, as it circumvents such scalability issues by only verifying the current system execution. Runtime Enforcement (RE) builds on RV by automating recovery procedures once a correctness violation is detected so as to mitigate or rectify the effects of the violation. We can therefore see Runtime enforcement as made of two parts: (i)Verification (Monitoring), and (ii) Recovery Actions.