Matching and ASM-based Choreographing of SemanticWeb Services Using F-Logic and Flora-2

In this dissertation, we mainly reincarnated the Semantic Web Service Choreography that had been abandoned for a decade in spite of its interesting utility, especially in today’s dominant web service-based applications. We took the original Abstract State Machine-based (ASM) choreography execution algorithm of Web Service Modeling Ontology (WSMO), determined its weaknesses and improved it so that it can be used effectively in the context of semantic web services. We then implemented a completely new choreography engine based on our improved algorithm. Our work has been done in two phases. First, we concentrated on the “capability” component of WSMO and used Frame Logic (F-Logic) to specify it. The new model allows very short but expressive descriptions of both goals and web service capabilities, which are then used by a matching engine to discover which web services can satisfy a given goal. The matching engine, using the meta-level F-logic inferencing capabilities of the underlying Flora-2 reasoner, is very efficient and has a very concise definition itself. In the second phase, we again used F-Logic for specifying ASM-based modeling of interactions between a requester of service and provider of service, also called choreography, of semantic web services described in conformance to WSMO. Our choreography execution engine, implemented in Flora-2, remains loyal to the parallelism and branching paradigms of ASMs (unlike previous implementations), and is based on our improved choreography execution algorithm, which has the following novelties over the original algorithm: (i) it introduces the concept of initial state in the execution of ASM and links it to the precondition of the goal, (ii) it introduces the concept of a final state in the execution of ASM and links it to the post-condition of the goal, (iii) it modifies the execution of ASM so that it stops when the final state conditions are satisfied by the current configuration of the machine, as opposed to stopping only when the machine has no more moves. As part of our work, we also developed a visual tool for specifying web service choreographies in Flora-2, provided a mapping between JSON and Flora-2 Web Service specification, and lastly defined a formal mapping between traditional ASMs and ontological ASMs that are the basis of choreography specifications and proved their equivalence, which was missing in the literature before. Keywords: Semantic web, web services, service matching, service choreography, abstract state machine, Flora-2, F-logic, web reasoning