Research Of Dynamic Service Replacement On Internetware

In recent years, with the development of Internet, the traditional component development methods can not meet the challenges of resource sharing and integration in an open, dynamic and difficult to control network environment. Service-oriented computing and architecture development method is based on the further development of components in the Internet environment. Web services technology as an important technology on service-oriented computing and architecture will enhance the component and the component assembly to the level of an independent service and service composition. However, with the increasing complexity of the users’needs,a separate service could not satisfy the requirement of user. Service composition technology combined single service to meet the needs of users of complexity. But, service composition mechanism is very complex, especially the dynamic service composition. Once the composition process is failure, some services have the same function should be find and then replace the failure service in short time. This can keep the consistency of the whole services composition. So, dynamic replacement technology has the actual value and should be focused when do research on service composition.This thesis analyzes the theories and techniques about the replace technology on services composition. Based on the analysis, we propose a semantic-based service classification algorithm. The algorithm to extracted the target service’s WSDL document semantic information, then extend the extracted semantic information through BabelNet synonyms dictionary. Through the extension information,we get a synonyms word set. Then, we get the similar functions services set through cluster algorithm. Before running the services composition, algorithm separate all services into small sets, can help reduce the number of the target collection when finding the same function services. The small collection can help to improve the efficiency of service discovery in the process of running a combination process. On the basis of semantic-based service classification algorithm, we establishment functional service community through solving problem on the number and on the sequence and the names of heterogeneous parameters. Then we get the service community table of the mapping situation between the parameters. This table also can give us a solution of the problem about the service heterogeneous parameters.The thesis also proposes a three-tier architecture based on dynamic replacement semantic classification algorithm and service community model. Model to achieve a dynamic service discovery, sort, select, Model gives us an objective standard to find service by combining the QoS evidence databases, and help us to achieve dynamic service replacement. Through simulation and experimental data, the model reduces the time to find service and improve the precision of the service, improve the service right replacement rate, shorten the time to replace the same function service.