Software QoS enhancement through self-adaptation and formal models

Abstract

Modern software operates in highly dynamic and often unpredictable environments that can degrade its quality of service. Therefore, it is increasingly important having systems able to adapt their behavior to the environment where they execute at any moment. Nevertheless, software with self-adaptive capabilities is difficult to develop. To make easier its development, different architectural frameworks have been proposed during the last years. A shared characteristic among most frameworks is that they define applications that make an internal use of models, which are analyzed to discover the configurations that better fit in the changing environments. In this context, this tutorial presents the current research advances on architectural frameworks for building self-adaptive software that meets its Quality of Service (QoS). We discuss architectures that use self-adaption to improve the QoS and whose adaptations are planned as a result of the analysis of formal models. We also describe a set of current research challenges that are still preventing the complete automatic control of dependable self-adaptive software.