Study On Distributed Web Services Discovery Mechanism

With the growing of Service-oriented computing and Web Services, more and more Web Services are becoming available on the Internet. The increasing number of Web Services demands for a scalable, flexible and reliable solution to discovery the most appropriate services. A distributed and dynamic service discovery mechanism becomes a key requirement for service requestors.According to the research trends in this area, the thesis is focused on distributed Web Services discovery problem and makes an in-depth study on how to find a service On the Internet by distributed infrastructure and semantic information.The main contributions of our work are discussed in the thesis in detail, and summarized as follows:1. A temporal description logic based service behavior description model is proposed.Based on analyzing the requirements of the semantic service behavior description, a temporal description logic (TDL) based semantic service behavior description and matching method is proposed. The TDL-based service behavior description model can describe temporal constraint of service behavior and ensure correctness of the service scheduling in process of web services discovery. Service discovery can make use of service behavior information through semantic description of TDL-based service temporal constraint.2. A structural similarity measure of web services is presented.Service description information is a kind of structured hierarchical data. The text or key based service matching can not distinguish structural differences of the service description. A novel service matching method is proposed to deal with service descriptions with deferent structures. The service matching method uses graph theory to measure similarity between deferent structures of services. 3. A distributed semantic web services locating model based on semantic overlay network is proposed.A novel peer-to-peer indexing system with related P2P registries is proposed to support the completely decentralized discovery capability. With the ontology encoding scheme, semantic service description is distributed into a distributed trie index on structured P2P network Requesters can lookup services with semantic requirements.To manage service advetisements, the presented system also extends the Plaxton mesh model to dynamic semantic overlay network to manage service advertisements. A Plaxton-like prefix routing to semantic routing model is extended and makes it possible to quickly identify the peers containing most likely matched services according to user requests. Peers can be clustered according to measurement of semantic similarity, and messages can be routed according to their semantic matching degree.4. A distributed service discovery mechanism based on distributed data mining technologies is proposedA distributed data mining algorithm build on P2P network is proposed to deal with distributed services. A service clustering algorithm is proposed to divide web services into difference classes. Then a distributed services classification algorithm matches the service requirement to special class to fulfill the service discovery.