A rule-based approach for dynamic adaptation of service flow

In this paper, we propose a rule-based dynamic approach for the online adaptation of service flow. The situations that may trigger the dynamic adaptation are analyzed and categorized. For each type of situation, a set of Event Condition Action (ECA) rules is devised to identify the problem and automatically select proper alternative services or subflows to remedy the situation. The alternatives are selected to retain the semantics of the original flow but comply with the new situation or QoS constraints. The scope of adaptation is deliberately arranged to minimize the impact and conserve the recovery cost. Preliminary implementation and evaluation results demonstrate that our techniques can successfully identify problematic situations and construct alternative flows to satisfy user requirements and QoS constraints with very low overhead. The generality and flexibility of our approach also enable the proposed techniques to be applicable to a wide range of applications.