Semantic Service Modeling: Enabling System Interoperability

1.1 Introduction The integration of software systems is a major challenge for companies today. Both external forces, such as business process integration, and internal forces, such as the move towards service-oriented architectures, put increasing pressure on software engineers to reuse and integrate existing system services, rather than building new systems from scratch. However, the lack of interoperability of software systems forms a major stumbling block for the integration of such services. Hence a lot of effort is currently being invested in standardization of service description languages and protocols for service interactions [2][9][12]. Unfortunately, these efforts mainly address what we call syntactic interoperability, with semantic interoperability just starting to be addressed in a number of initiatives (see next section). These initiatives propose semantically-rich service models, definition of mappings among these models, and runtime mediation based on the defined mappings. Interoperability is the capability of different systems to use each other's services effectively. It is about sharing functionality and information between systems at different levels, e.g., between physical devices, software applications, business units within one organization, or between different organizations. Interoperability implies that systems are able to interact (i.e., exchange messages), read and understand each other's messages, and share the same expectations about the effect of the message exchange. In this paper we analyze and define in detail what it means for software systems to be interoperable. We identify three different levels of interoperability – the syntactic, semantic and pragmatic level – and define the requirements for assessing interoperability at each of these levels. We propose a method for formally verifying the semantic and pragmatic interoperability of a number of systems,