This paper presents OWLS-CSM framework, an OWL-S Service
Keywords: web service discovery, subsumption hierarchy, OWL-S service profile, taxonomical similarity, milestone
IntroductionThe increasing use of web services in cloud environment [1] has raised new vital challenges, such as the finding, the most suitable web services that satisfy users' requirements.Adding annotations to the services in cloud environment, which then called semantic web services (SWS in short) [2], aim at making web services machine understandable and useapparent, utilizing semantic web technologies for Web service annotation and processing. The idea is to provide ontology-based descriptions of web services that could be processed by ontology reasoning tools. Thus, intelligent agents can be able to automatically understand what a web service offers and what it requires in order to execute a task and facilitates selecting the suitable service that meets user's needs.In this paper, we propose an OWL-S Cloud Services Matchmaker (OWLS-CSM) framework, in which user can find the suitable service which meets his needs. In OWLS-CSM, a conceptual model for semantic web services is adopted and WSMO Discovery Framework [3] (WSMO-DF) is modified and utilized for Web service discovery. An OWL-S ontology is built and semantic descriptions are expressed as instances of the profile concept of the OWL-S Service Profile (SP). Semantic descriptions cover functional properties -which can be matched latersuch as inputs, outputs, precondition, effects and nonfunctional properties, in order to capture fast a set of candidate Web services that meet user's needs.This rest of paper is organized as follows. Section 2 presents the related literature includes the conceptions we use in our study. Section 3 introduces the theory of service profile similarities and metrics used in the proposed framework. It shows the three levels of similarities and their algorithms as well. Section 4 introduces the environment that OWLS-CSM works in and analysis the obtained results and compare them with similar previous works in order to show the efficiency of our framework. Finally, there is a section for conclusions and future work.