Study On Fuzzy Logic-based Privacy-aware Web Service Selection

In recent years, with the wide application of Web service technology, a large number of Web services have appeared on the Internet. These services provide consumers with various functions through the Web platform, which form the basis of large-scale service-oriented computing. Due to the constant competition among service developers, services with the same or similar functions are increasing rapidly, making service selection based on the user's requirements becoming a even more challenging issue than ever. So far, a number of Web service discovery and selection methods have been proposed to address the above issue. Most of them are based on quality of service (QoS) evaluation or trust evaluation. However, the previous approaches rarely take into account the issue of user privacy protection. In a service selection scenario,the user usually needs to provide its personal data to the service provider, such as name, cell phone number and so on. These data are probably concerned with the user's personal privacy, thus service consumption may cause the exposure and abuse of the user's personal privacy. Since there often lacks an available privacy data protection agreement between a user and a service provider,how to prevent the user's privacy from exposure and abuse becomes an important issue.This paper mainly makes the following contributions: 1) A flexible expression method is proposed for the definition of the privacy policies of service providers and the privacy requirements of users, allowing users and service providers to describe their privacy preference and privacy policies with natural language terms; 2) Aiming at protecting the user's privacy in service selection, a set of fuzzy logic reasoning rules are developed via combining the reputation of services, the privacy policies of services, and the user's privacy preferences. Based on the fuzzy logic reasoning rules,the privacy cost of each service candidate can be calculated, and the service with the smallest privacy cost can be identified and recommended to the service user; 3) A privacy-aware service composition method is proposed via adding the user's privacy protection requirements to the service composition. The method models the privacy-aware service composition issue as a multi-objective optimization problem on a graph, and then transforms the multi-objective optimization problem service composition into the optimal path finding problem, which can be tackled using a modified Dijkstra algorithm.To demonstrate the effectiveness of the proposed methods, we not only used a number of examples to illustrate the method, but also developed a privacy-aware service selection prototype system, and also some experiments were conducted to evaluate the performance of the proposed methods.