Self-adaptive Method Based on Software Architecture by Inspecting Uncertainty

Abstract

A recent common approach to monitor and adapt system behavior at runtime is to decouple one or more external modules and self-adaptive mechanisms from the target system. The non-invasive manners have the main advantage of realizing separation of concerns. However, some uncertainty aspects emerge while utilizing these separate control units. The unanticipated inherence and complexity of upcoming services and applications make proactive self-adaptation essential. In this work, the specification of software architecture is extended using CHAM (Chemical Abstract Machine) by inspecting uncertainty. Software architecture guides the topology of the constituent computational elements (such as components and connectors) of the system under consideration. The proposed self-adaptive model is novel as it leverages standard software architecture models, and quantifies behaviors of the system in terms of relevant architectural elements. Experiment results show the effectiveness of the proposed method.