QoS-Aware Web Service Semantic Selection Based on Preferences

Abstract

The emergence of service oriented computing brought major research challenges in the process of discovery and selection of web services. Key to success in obtaining the best fitting web service out of a multitude of alternatives is finding the service that both satisfies the requester's functional needs and non-functional requirements. Clients should also be able to indicate how to make trade-offs when some of thir requirements cannot be met. The ability to capture trade-off preferences is critical for selecting the best fitting web service. In this paper, we address the problem of expressing the requester's preferences in a semantic context. We analyze the existing QoS ontologies and choose to extend the OWL-Q ontology to capture trade-off preferences expressed using QoSPref, our own conditional lexicographic approach for QoS preference specification. We extend the QoSSpec specification facet of OWL-Q by adding a new class QoSSelectionWithTradeoffs, that offers the possibility of describing the list of preferences and trade-offs similar to our QoSPref notation, thus offering the possibility of using our algorithm for ranking web service alternatives.