Trusted Web Service Selection Approach Considering Uncertainty Of QoS Data

Web service as Internetware, which is supported by software componet technology,supports collaborative operation between different applications in the network and onlineevolution of distributed software systems, has become a new kind of web application form.Web services deployed on the Internet can be accessed and integrated, these servicescontinue to enrich and constitute a huge library of standard componets. Faced with a largenumber of same or similar function Web services, and how to select avalible, reliable andtrusted Web services for users has become a key issue. Web service selection results arenot only directly related to whether services can be called successfully, but also have acritical influence on the quality of service, QoS distinguishes the differences well of Webservices, trusted service selection based on QoS has become a hot research. This paperfocuses on Web services’ trust, QoS data uncertainty and other key technology issues inWeb service selection in-depht study.Firstly, an approach of Web service trust evaluation conbined with subjective trustand objective trust based on the duality of trust is given. Moreover, a trusted recommenderdiscovery algorithm based on trust tree is proposed after comprehensive analyzing theeffect of recommenders’ reliability to trusted Web service selection. The algorithm makesuse of trust relationshiop network formed by entities and constructs a trust tree, gives thetrust tree’s data search mechanism based on tree’s breadth-first traversal.The algorithm canlocate the trusted recommenders accurately improving Web services’ trust, and as well asovercome the low efficiency of trust data searching in trust network.Secondly, a service selection approach considering the uncertainty of QoS data wasproposed on the basis of comprehensive analysis of QoS data uncertainty connotation. Thedifficulties of accurate and reasonable QoS constraints definations are taken into accountbefore Web service selection; three eigenvalues expectation, entropy and hyper entropyare used to predict the distribution of QoS data of Web service and to analyze theuncertainty of QoS data. By using the QoS constraints defined by users, QoS attributecloud model is built to compute the QoS data uncertainty for filtering redundant services while extracting reliable services. Then, the QoS attribute feedback similarity is calculatedas the weights for QoS aggregation, QoS evaluations and the numbers of service users areintegrated for computing service recommendation degree which is the standard for servicesorting and selection, thus ensures the sufficiency and objectivity of selected services.Finally, in order to verify the feasibility and effectiveness of proposed trustrecommender discovery algorithm based on trust tree and service selection approachconsidering the QoS data uncertainty, IWebService service selection platform is designedand developed. A Web service selection instance based on actual application scenarios isdemonstrated in IWebService, and the prososed algoritm and approach are analyzed andverified combined with selected results.