Web Services Discovery with Rough Sets

Web services are emerging as a major technology for building service-oriented distributed systems. Potentially, various resources on the Internet can be virtualized as Web services for a wider use by their communities. Service discovery becomes an issue of vital importance for Web services applications. This article presents ROSSE, a Rough Sets based Search Engine for Web service discovery. One salient feature of ROSSE lies in its capability to deal with uncertainty of service properties when matching services. A use case is presented to demonstrate the use of ROSSE for discovery of car services. ROSSE is evaluated in terms of its accuracy and efficiency in service discovery. INtrODUctION Web services are emerging as a major technology for developing service-oriented distributed systems. Potentially, many resources on the Internet or the World Wide Web can be virtualized as services for a wider use by their communities. Service discovery becomes an issue of vital importance for Web service applications. As shown in Figure 1, discovered services can either be used by Web service applications or they can be composed into composite services using workflow languages such as BPEL4WS (Andrews Curbera, Dholakia, Goland, Klein, Leymann et al., 2003). UDDI (Universal Description, Discovery and Integration, http://www.uddi.org) has been