Towards the automation of the service composition process: case study and prototype implementations

Service Composition has been a challenging research area for many years. One of the key ideas in this area is the matchmaking (at the semantic level) of requested services and the portfolio of services registered in a given service repository. Accordingly, a composite service can be generated by selecting a set of existing service components that partially match the requested service, and composing these services. In this paper we introduce a general framework for the Service Composition process and our efforts to automate this process. This framework exploits the semantic annotation of services and their parameters. The main annotations we focus on are service inputs, outputs, goals, preconditions, effects, and non-functional properties. Matchmaking aims at making sure that the selected service components are capable of interacting with each other, while the composition aims at orchestrating the selected service components to fulfil the goals of the service request. The framework takes care of service requests from both experienced service developers as well as ordinary end-users who are not necessarily familiar with service concepts and platforms. Three main prototype platforms have been developed to experiment with the framework.