The Research And Implementation Of Web Service Selection Based On The Auction Theory

With the rapid development of network technology and e-commerce, the Web Services with single function are often unable to meet the increasingly complex business requirements. People pay more and more attention on how to effectively integrate services distributed in the network and realize powerful composition services. However, because of the increment of the Internet size and the dynamic network environment, service providers tend to inflate its services QoS in order to maximize their benefits. Therefore, the traditional method of Web service composition has been unable to find a composition-service to meet the needs of users in the best possible time. To solve the above problems, we propose a Vickrey auction based Web service composition method, which not only ensures that the Web service providers do not inflate its service QoS, but also considers the global QoS constraints and the local QoS constraints of single node in the conposition service process. The main work is as follows:1. A Web service composition model based on auction theoryThe model includes:an abstract service composition framework (Auction-Based Service Composition Framework, ABCF) and QoS composition model.The ABCF consists of three participants, two function modules and a service registration center.Participants are User, Broker and Seller. User sends a service composition request to the Broker.Broker is responsible for the choice of service, and return the result to the User. Seller to Provide Web service with QoS attributes. Function modules include:auction module and global optimization module. Auction module is responsible for the Web service auction and local service selection. Global optimization module is responsible for adjusting the QoS constraint. The service registry provides the function to publish and query services.User not only pay attention to the functional requirements of service when chooseing services, but also the QoS attribute. The quality can not be determined directly by the QoS value of Web services because of the range of QoSs are different.Therefore, the paper propose the QoS model.It firstly normalizes the QoS to the same range, then calculates each service utility with the weighted sum approach, last the calculation methods of composite service QoS is given.2. Proposed two reverse Vickrey auction based service selection algorithm.Based on the above auction model, this paper puts forward a reverse Vickrey auction based service selection algorithm (RVBSSA). RVBSSA includes local selection and global adjustment phase.In local selection stage, choose K (K>0) specific Web service with maximun utility for each abstract service node of the combination process with reverse Vickrey auction. In the global adjustment stage, firstly it checks whether the global QoS of the composition service getted by local selection meet the global QoS constraints. If meet, the composition service is the optimal solution, otherwise, RVBSSA need to adjustment the QoS value which not meet the constraints by descend order according to the user’s preferences. The paper puts forward a reverse Vickrey Auction negotiation algorithm (RVBSSA2) which based on RVBSSA algorithm. It aims to incentive that Seller would provide more realistic QoS attribute values through auction negotiation. Through the experimental analysis between the two algorithms and integer programming (IP) algorithm, our algorithm’s execution time is far less than the IP algorithm while the performance is close to the IP algorithm.3.Implementation of Prototype SystemBased on the framework of ABCF service composition, the paper simulate the composition process using the NetBeans tool, realizes the function of each module, achieve Web service registration center wiht the MySql tool. The operation of the prototype system indicates that the system can meet the customers’requirements and find out the approximate optimal combination of the Web service quickly.