Towards Web services tagging by similarity detection

A challenge for the Smart Internet will be the automated tagging of equivalent or similar services, both in terms of domain semantics and service protocols, in support of efficient discovery and selection of relevant alternative services for the current matters of concern. Code similarity detection is an established technique that can be brought to bear on this problem if service descriptions can be partitioned into appropriate units for comparison. Unfortunately, specifications written in Web Service Description Language (WSDL) are poorly structured for this purpose, with relevant information for each service operation scattered widely over WSDL service descriptions. In this work we describe a first step in leveraging code similarity techniques to identify and tag similarities in WSDL descriptions of web service operations. Using source transformation techniques, we describe a method for reorganizing WSDL descriptions such that they are both more human readable and better suited to analysis by similarity detection tools. We demonstrate our method by the automated identification and grouping of similar service operations using clone detection in two example WSDL systems. © 2010 Springer-Verlag Berlin Heidelberg.