Discovering semantic Web services via advanced graph-based matching

One of the main advantages of Web services is that they can be composed into more complex processes in order to achieve a given business goal. However, such potentiality cannot be fully exploited until suitable methods and techniques allowing us to enable automatic discovery of composed processes are provided. Indeed, nowadays service discovery still focuses on matching atomic services by typically checking the similarity of functional parameters, such as inputs and outputs. However, a more profitable process discovering can be reached if both internal structure and component services are taken into account. Based on this main intuition, in this paper we describe a method for discovering composite OWL-S processes that founds on the following main contributions: (i) proposing a graph-based representation of composite OWL-S processes; and (ii) introducing an algorithm that matches over such (graph-based) representations and computes their degree of matching via combining the similarity of the atomic services they comprise and the similarity of the control flow among them. Finally, as another contribution of our research, we conducted a comprehensive experimental campaign where we tested our proposed algorithm by deriving insightful trade-offs of benefits and limitations of the overall framework for discovering Semantic Web services. © 2011 IEEE.