PASCAL: A Protocol-Based Approach for Service Composition and Dependable Optimization

Service composition is a promising technique for developing applications especially for those across multiple organizations. The dependability of composite services, however, is difficult to be guaranteed due to the distributed, dynamic and autonomous characteristics of service domains. In this paper, we address this problem by proposing a novel approach-PASCAL (Protocol-based Approach for Web Service Composition And Dependable Optimization). We initially propose a protocol based service composition system architecture supporting online monitoring of services interaction and dynamic replacing of service. Then, we design the dependable optimization mechanism based on the service and dependencies monitoring. Furthermore, we design an algorithm to adjust the configuration adaptively base on policies and protocols to keep the composite service dependable while service failure appears. Finally we implement the PASCAL approach based on XService suit and evaluate the proposed mechanism through comprehensive experiments and achieve improved results.