Research On Web Service Resource Selection Technology In Cloud Environment

Nowadays the Service Oriented Architecture(SOA)become the first choice for enterprise,it makes the Web services' numbers continuously rapid growth,along with the rising number of users the Web services' demands is expanding.how to make Web selection more accurate and quickly in a huge Web services resource is a tough problem in the project.Although the existing service recommendation systems that can make recommended,but the aspect in service recommendations' real-time and accuracy was not considerable enough,there are many users' requests cannot consider the conditions of time changed and in dynamic service recommendation,also there is not involved the interrelationship such as service response time?service recommendation effectiveness and service accuracy,so it directly lead to the selected contents cannot get users' heart and failed to meet the users' service demands.In order to improve the response time and service recommendation accuracy,how to adopt the Web service selection method in the dynamic service selection to meet the users' service demand as the entry point,considered from two aspects of real time capability and effectiveness of the service to begin the study,the paper's main content as following:First of all,through extended the Web services description language,in this way to redefine the port to redefine the Web service,then introduce the new description as a new description into the uncertain reasoning net of Bayesian probability,and to make the bayesian network in dynamic change time to divide the time sheet and restrict the recommendation service time.Secondly,in bayesian selection service model by determining three parameters,including the time slice parameter,extension service response time,the service semantics parameters and so on to create a new DBN,it is using for screening the users' service informations and ensureing the users' servicerequests can be selected in a response time.At last,because of the selection of Web services has the problems in cold start and new users,the paper based on the collaborative filtering method the paper use community affect to calculate K points then making K-Means clustering to found community,and manage the users by the community classification management,when the mode faced with a new user or have no information to make bayesian dynamic selecting conditional failure,the service recommendation module will using the management information to calculate its similarity user data to complete service resources recommended.In conclusion,DBN selection technology based on WSDL extended,no matter the users' service was very large nor the cold start and new users,the new model could fastly and accurate select users' service resource.Experimental result demonstrates,based on community found that the way of data preprocessing in both the time and the speed should be higher than that of pure K-Mean scheme according to the processing mode,so it is a better way to assist the selection.In BN to use the Bayesian method which has time analysis to predict the users' service,by the time limit cognitive way to improve the accuracy of BN service selection,and ensure the real-time and effectiveness of the service selection.After in heilongjiang province road transport pipe of Yun Zheng instance experiments in the test system.Found that use this way to get more real and more accurate recommendations' result,and saved more business processing time improve the efficiency of business processing,finally from the project evaluation points it has the good satisfaction.