Runtime Failure Prevention and Reaction

International audienceThis chapter describes how to use in-the-field runtime techniques to improve the dependability of software systems. In particular, we first present an overall vision of the problem of ensuring highly-dependable behaviours at runtime based on the concept of autonomic monitor, and then we present the two families of relevant approaches for this purpose. First, we present techniques related to runtime enforcement that can prevent the system producing bad behaviours. Second, we describe healing techniques that can detect if the system has produced a bad behaviour and react to the situation accordingly (e.g., moving the system back to a correct state)