The Research Of Trust Model Based On Application Layer Multicast For Streaming Media

With the rapid development of streaming media, how to make the transportation of streaming media in network secure, stable, high-speed becomes the hot spot of research. Application layer multicast is used widely, because it doesn't need additional hardware establishment, modify network protocols and facilitate the deployment of streaming media applications on the Internet.However, due to the opening and anonymity of the application layer multicast, the large maintenance cost, difficult management, fraud of malicious nodes can be seen everywhere.The root reason of these problems is the nodes of the application layer multicast lack of effective restraint mechanism and lack of trust between nodes. Trust model is the most important method for solving the security issues of the application layer multicast. However, the existing trust model has many disadvantages such as unilateralism of calculating trust, inaccuracy and malicious behavior can not be effectively suppressed. In addition, how to enable the trust model to run stably in the dynamic network of the application layer multicast is a difficult problem too. On the basis of basic idea of the existing trust models, this paper proposes an improved trust model based on reputation according to the analysis of core technologies. This model divides the reputation value into direct evaluation of reputation and indirect evaluation of reputation and calculates the reputation value according to their different weight. For the direct evaluation of reputation, the punishment mechanism which warns the behavior of malicious nodes is introduced For the indirect evaluation of reputation, considering the recommended value and the credibility of recommendation. The credibility of recommendation will affect the proportion of the recommended nodes in the reputation value. For the low recommended credibility nodes, the recommendation weights less impact on the calculation results, thereby helping to avoid the adverse effects of malicious recommended. And this paper proposes the randomly extraction algorithm based on credibility of recommendation. The main function of the algorithm is when the recommended nodes are excessive, extracting randomly quantitative values using this algorithm to avoid the rate of the node communication will decline significantly because of a large amount of computation.In order to make the trust model running stably and effectively in the application layer multicast, this paper proposes the algorithm idea of building a stable application layer multicast and improving the stability of application layer multicast.At last, the paper designs trust model architecture based on the application layer multicast. It describes the total process from nodes joining a multicast tree, communicating under the protection of trust model and leaving multicast tree. This architecture adjusts to the application of stream media technology because of great improvement of stability and security in its transmission.Simulation experiments demonstrates that compared with the traditional ones, the proposed trust model has better quality in resisting against the malicious behavior in application layer multicast and better solution for the safety and stability of the application layer multicast...