Research On The Key Technologies In The Security Access Of Web Service

Web service has an important effect on user experience and enterprise applications.Because of the dynamic performance of the open network, the interaction among the participants in Web service access is always irregular. At present, the researches on the process of security Web service access have fellow problems. (1)The traditional Semantic-based Web service matching algorithms couldn’t deal with this situation when it had fuzzy semantic concepts; (2) The trust evaluation to recommended paths of Web service was still unreasonable; (3) Current description and optimizing mechanism of Web security policy weakened the descriptive function of security assertions.Aimed at these problems, this paper relies on 863 projects to research the three key technologies in the security access of Web service:the matching of Web service, the trust evaluation to recommended paths and the optimizing mechanism of Web security policy. In this paper, the security problems of Web service were analyzed so that the matching accuracy could be improved, the multi-trust paths could be optimal aggregated, and the conflict of the policy could be minimized. The main work and achievements are as follows.1. A Web service matching algorithm based on dynamic trust-based semantic library was proposed to deal with this situation when it had fuzzy semantic concepts. A dynamic semantic library was created by extracting semantic information from the service description recorded in the trusted entities. In the library, the fuzzy semantic concepts were replaced by that extracted from the recommended service description. The semantic similarity was calculated at last among the vectors of the service request and the service description. The experiment has shown that the proposed algorithm can cover the shortage of the traditional semantic-based matching algorithms. The matching result in fuzzy semantic environment can have a 91.5 percent integrating degree with that in exact semantic situation.2. A new path selection and aggregation algorithm based on Grey & Hidden Markov Model (G-HMM) was proposed to solve the problem which the Web service access had in the prediction and selection of multi-trust paths. In order to filter the malicious path, the model analyzed all the entities in the system using the grey clusters theory, and predicted the acceptation and grey state which the path had by the forecast of Grey Markov Model. The Hidden Markov Model was given to evaluate the optimal path finally. Compared to current algorithm, the proposed model can filter the malicious recommend entities and promote the effectiveness of the recommendation system. The proposed can promote 0.7 to 1.8 percent more than the predominant algorithm on equal conditions.3. A security policy optimizing algorithm based on minimum irrelevant attributes set (MIAS) was proposed to settle the situation that current security policy optimizing algorithm laid emphasis on the concepts, causing, and classification and so on, yet neglected the description of the security assertion. The necessary attributes for Web service access was extracted from the Web request. The MIAS was constructed by removing the correlation among the attributes and deleting the invalid attributes. The assertions were described by MIAS and the policy was constructed by the assertion in the process of logic "AND". Based on the security policies, the policy optimizing mechanism was proposed at last. Compared to the classical policy conflict optimizing algorithm, the proposed has better detection performance. The maximum detection time can decrease 48 percent in the same testing environment.