A Service Selection Workflow for Composition Using Correlation and Route Optimization

Web services have been widely accepted as a server-side technology and are the key component to cloud computing. How to effectively compose a service from a number of services is a still an issue. The current solutions are either too specialized or too complex. In this paper we present a workflow that models service composition as a service graph where each node is an individual service. Then the problem of service composition becomes the problem of path optimization. In our approach the services are first evaluated and ranked according to their quality of services. The highly ranked services are selected to construct a service graph determined by the correlation values among them. Finally, several service selection strategies are developed according to different correlation and path optimization policies. These mechanisms are further evaluated and validated by simulations. The experiments show the proposed methods are feasible and not complex.