Research On Skyline-based Distributed Service Selection And Method Of Dynamic Maintenance

With the development of SOA, more and more distributed systems put into use, which makes the scale of web service and the complexity of Web environment keep increasing, that cause many traditional service selection methods lose its original performance. Thus the article points to Web service selection problem of distributed dynamic environment do a through research and come up with a kind of algorithm that solves the service selection problems of distributed system. Then it puts forward a self-maintenance mechanism in dynamic web environment, which making the service selection algorithm remains a better performance when dynamic changes occurred.First, in the case of a comprehensive analysis of the current service selection algorithm,the Skyline computation in the field of database is introduced into the service selection by thinking about the distributed system efficiency's problems. After the QoS and semantic of Web service have been modeled, a QoS-maintenance distributed Skyline service selection algorithm (FSSDS) based on feedback was proposed, which can fastly selecting a current set of optimal service by using the Skyline computation in each distributed node. Then FSSDS gets the Skyline sets together, and references ontology for the level of semantic matching.FSSDS select the global optimal set of services by using this hybrid mechanism to satisfy user's functional and non-functional requirements.Then,taking the dynamic Web services environment into account, a self-maintenance model (DSM) based on FSSDS algorithm was proposed. And then, a maintenance principle for DSM model was proposed. Putting these together, the paper puts forward a DSM-Skmaintain algorithm which can judge in every change of service, then perform different maintenance strategy on different types of change to make the current node always maintain a set of optimal service.At last, combining FSSDS with DSM Skmaintain, a series of simulate experiment was carried out for validating and analyzing.