Two-phase web service discovery based on rich functional descriptions

Abstract

Discovery is a central reasoning task in service-oriented architectures, concerned with detecting Web services that are usable for solving a given request. This paper presents two extensions in continuation of previous works towards goal-based Web service discovery with sophisticated semantic matchmaking. At first, we distinguish goal templates as generic objective descriptions and goal instances that denote concrete requests as an instantiation of a goal template. Secondly, we formally describe requested and provided functionalities on the level of state transitions that denote executions of Web services, respectively solutions for goals. Upon this, we specify a two-phase discovery procedure along with semantic matchmaking techniques that allow to accurately determine the usability of a Web service. The techniques are defined in the Abstract State Space model that supports several languages for describing Web services. © Springer-Verlag Berlin Heidelberg 2007.