Indicates That The Service Calculated Risk Information And Access To Research

Service computing uses service as the basic elements of development and application, and offers a new computing model for Enterprise Application Integration heterogeneous applications of cross-platform, loose coupling. Web Service technology uses WSDL (Web Services Description Language), SOAP (Simple Object Access Protocol), and other standards and protocols, solves the heterogeneous distributed Computing, code and data reuse. It has highly interoperability, platform independence, simplicity, universality and loose coupling characteristics, and has become the mainstream Service Computing implementation technology. However, Web services only provide a single function, can’t satisfy different users’complicated application requirement, so it’s necessary to make existing Web services be dynamically found and organic synthesized according to certain rules, then assembly become a web service which is value-added, bigger size, and have internal process logic. It is the service composition problem.On the other hand, with the popularity of Web services, the accessible Web service on the Internet dramatically increased, leading to a large number of the same function or similar Web services appear. Therefore, on the selection of Web Service, it is necessary to consider price, execution time, reliability and other non functional attributes of the Web services, namely QoS (Quality of Service). Open and dynamic environments and the diversity of QoS lead to the uncertain description about performances of Web services, using Web service with uncertain QoS of to structure composition service, will inevitably lead to composition service users face to the risk of QoS demand cannot be satisfied. Therefore, this thesis aim at the composition service QoS risk sources, representation, quantitative, identification, control for system research, so as to improve the composition of services ability that meet users QoS demand and reduce the composition service risk appear very important.This thesis mainly researches the representation and acquisition of composition service QoS risk information. To manage the risk of service, it is necessary to obtain stochastic character of QoS that is represented as accurate probability distribution of QoS. This thesis presents an approach to estimate probability distribution of Web service QoS in the case of small number of samples. Using max entropy principle, the analytical formula of the probability density function can be obtained by transforming the probability distribution estimation problem into an optimal problem with constraints obtained from sampling QoS data. Then an algorithm to estimate parameters of the probability density function is designed. At last, aim at the workflow in the process library which can act as a composition Web service object, designed a prototype system for estimate probability distribution of composite Web service QoS. The system adopts C/S structure, uses Java and matlab language to implement. First of all, users load a workflow from the process library, then input QoS small sample set for describing a single Web service (response time, reliability, cost and so on). Finally, it analyses parallel, branch, sequence and loop structures, using the maximum entropy method for calculating QoS probability distribution curve of the entire workflow. We test the effectiveness and reasonable of the maximum entropy method for calculating the probability distribution curve of Web service composition QoS and can be accurate to provide information service for risk management by extensive experiments.