QoS Aware Adaptive Service Composition

Abstract

In current business scenario the business processes of enterprises are service oriented. The services of various applications are advertised in service registry using Web Service Description Language description for service selection and composition. Service composition process integrates services based on Business Process Execution Language description of composite service and executed by the orchestration coordinator. As all compliant services description are included in the description, the service selection and execution consumes more time users need select the service form set of complaint services. This study proposes an adaptive service composition methodology over extended web service Quality Aware Service Integration architecture (QASI). The proposed service composition methodology reads the composite service description and constructs a composition tree. In each service node of composition tree a service is replaced by one that satisfies functional and QoS requirement. During execution of composite service or binding of service, the parameters like server maximum capacity, current load on server and expected time for completing current load for service at each level of composition tree are analyzed. An alternative service will be allocated at particular level to improve the overall response time of composite service. The QASI architecture provides a repository to store composition tree as symbolic notations to ensure repeatability. Experimental results indicates selection of best service and better response time by proposed methodology using QASI architecture compared to contemporary web service composition methodology using static BPEL description.