The Research Of Support Correlation-service With Web Service Selection

With the development of service-oriented computing technology, the Internet provides moreand more services which are the same or similar functional, but with different non-functionalproperties. When building a composite service, the users have to face to the massive choosableservices. The key to Web service composition is that how to select the services quickly, whichsatisfies the user’ needs.Existing Web service selection methods generally consider that “candidate services areindependent of each other, the quality of a service is independent of other services” and “selectingservices successfully for the first time” as a hypothesis. But in the actual service selection process,there are some relationships between services, such as business association, quality, etc. Theserelationships affect the quality of the composite service. This paper mainly aims at studying theservice relationship and the influence of relationship on the composite service QoS. In the servicecomputing environment, some Web services can’t meet the needs of users initially, but which canbe achieved by negotiation. In order to achieve equitable and maximize of both participation, thispaper studies the specific methods of service negotiation, and also proposes the method to solve theconflict between the service providers and service requesters.Dealing with the service relationship, this paper analysis three business relationships,including statistical correlation, business correlation and interface correlation. On the basis of this,this paper proposes a novel model for Web service description. This model supports therelationship between services and based on this model, this paper provides a model to calculate thesimilarity of services and a method to discover correlation service collection. On the basis ofcomposite services QoS models which support the service correlation, this paper turns the Webservice selection problem of support correlated-service into a mixed integer-programming model inorder to solve it.In the real application environment, there are existing two situations: one-to-many and many-to-many models between service users and service providers. At present, most research focus onone-to-many model, namely, a service consumer and multiple service providers, which ignore thecase of multiple service requesters. However, this paper considers the case of multiple negotiationparticipants, which is based on Agent technology, and uses the Skyline method to eliminate redundant service providers. According to the number of negotiation participants, evaluating prosand cons of the negotiation environment, and then based on the negotiation environment ofcomprehensive analysis and the attitude of participants, using signed SLA agreement to completethe service negotiation process.