Optimizing monitoring requirements in self-adaptive systems

Abstract

Monitoring the system environment is a key functionality of a self-adaptive system. Monitoring requirements denote the information a self-adaptive system has to capture at runtime to decide upon whether an adaptation action has to be taken. The identification of monitoring requirements is a complex task which can easily lead to redundancy and uselessness in the set of information to monitor and this, consequently, means unjustified instalment of monitoring infrastructure and extra processing time. In this paper, we study the optimization of monitoring requirements. We discuss the case of contextual goal model, which is a requirements model that weaves between variability of goals (functional and non-functional requirements) and variability of context (monitoring requirements) and is meant to be used for modelling mobile and self-adaptive systems requirements. We provide automated analysis -based on a SAT-solver- to process a contextual goal model and find a reduced set of contextual information to monitor guaranteeing that this reduction does not sacrifice the system ability of taking correct adaptation decisions when fulfilling its requirements. © 2012 Springer-Verlag Berlin Heidelberg.