Quality Attributes Assessment in Self-Adaptive Systems: An Empirical Evaluation

Self-adaptive systems are capable to monitor themselves and the context surrounding them, detect changes, and decide how to react to unexpected conditions with minimal human supervision at runtime. One of the challenges behind the development of self-adaptive systems is to handle the decision-making process during the analysis of trade off points between multiple quality attributes (QA). In Software Engineering, a widely-accepted method for evaluating software architectures QA goals is the Architecture Tradeoff Analysis Method (ATAM). However, few studies report on tradeoff analysis of QA in self-adaptive systems, despite its importance. Therefore, in this article, we proposed an adapted version of the ATAM to handle the particularities of self-adaptive systems. We employed the UPPAAL SMC to assist in the analysis of a set of QA. To assess whether the proposed adaptation is feasible in practice, we carried out an empirical evaluation by running the adapted ATAM in a self-adaptive system developed following the MAPE-K model. The yielded results indicate that the adapted ATAM could support the design decision process in self-adaptive systems thatuse the MAPE-K model