One of the most powerful feature of Web Services is that it allows combining existing services to deliver value added service. Unavailability of any service in the composition, due to various disturbances, during the execution time will stall the whole composition until the failed service has been replaced by another service with similar functionality. Selecting a new service randomly at the run-time may cost a lot in terms of response time, network bandwidth, etc. In this paper we propose a framework to reduce the cost, hence improve the response time of a Web Service composition.
展开▼