Multi-Staged Quality Assurance for Self-Adaptive Systems

The emerging approach to tackle the increasing complexity of today´s software systems is the use of self-adaptation techniques. Most often, self-adaptation is introduced in terms of externalized adaptation rules (e.g. event-condition-action rules). Modeling and implementing adaptation rules introduces an additional amount of complexity that potentially results in erroneous system specification models. As a consequence, quality analysis and assurance must be considered early in the development cycle of self-adaptive systems. In this paper, we propose an efficient quality assurance approach for self-adaptive systems in terms of an integrated modeling and analysis approach, which is capable to identify errors (e.g. conflicting adaptation rules) in the set of adaptation rules early in the design process. Based on Adapt Cases, a high-level design modeling language for self-adaptive systems, we describe a multi-staged model-checking approach and address the problem of state space explosion by considering properties that are specific for adaptation modeling.