Research On Discovery Mechanism For Semantic Web Service

As the research on the technology of web service develops, semantic web service was proposed for intelligentizing web service. Semantic web service is the integration of web service and semantic web, whose goal is to make web services understandable by computers to support automation of discovery, selection, negotiation, composition and execution of web service.Service discovery is significant for the whole lifecycle of semantic web service. Only finding the proper web service, can we perform successive stages of automation of web service. The traditional web service discovery is mainly based on centralized registry, such as UDDI. On one hand it can not support dynamic and complex queries well and suffers single point of failure and performance bottleneck; on the other hand, keywords matching which lacks semantic information of services can not guarantee the accuracy and quality of discovery. This paper is focused on how to find out semantic web services effectively and efficiently in distributed environment, in which a web service discovery mechanism on P2P network is suggested.First, this paper gives a creation algorithm of characteristic vector of semantic web services based on ontology partitioning techonolgy, so that the distributed web services belonging to the same concept groups are congregated together. In order to manage these clutered services, the prefix tree-based semantic web service discovery mechanism is proposed. In this system, peers are clustered into groups by their concepts groups and form a structured semantic P2P network presenting a toplogy of prefix trees. Prefix routing mechanism in the system makes full use of semantic information of services and has efficient semantic routing capabilities. Then the adaptive manner and load balancing strategy of the system are explained, which enhance the stability and usability of the network. The simulation experiment and PISOMWare application show that this system has stable routing hops, efficient discovery ability and it is also self-adaptive and load balancing, which is suitable for the large-scale ontology based distributed Web applications.