Improving Web Service Discovery by a Functional and Structural Approach

Abstract

Service oriented Architecture (SOA) has been widely used in service computing applications and this fact has been encouraged the publication over the Web in a Web Service format. Whereas the number of Web services published on the Web is growing up, discovery techniques must be improved so as to retrieve more desirable services. Nowadays, the most commonly used technique is semantic filters based on ontological concepts. However, such mechanisms can leave out some important Web services of the matching process, because of their structural relationship not mentioned in an ontology. In order to overcome such problems, some authors have proposed a hybrid approach to combine traditional syntactic and semantic approaches. These proposals remain restricted especially with respect to complexity, precision and time of execution, thus making such solutions in most of cases unfeasible. In this paper, we combine semantic filters based on functional properties with a structural approach, analyzing each neighbor relationship in an ontology. The results showed a considerable improvement in terms of performance and a complexity reduction with respect to other existing techniques. Furthermore, we implement a tool called OWLS-S Discovery in order to simplify the use of our approach by developers.