Research And Application Of Dynamic Trust Quantitative Mechanisms Based On Behaviors

Dynamic trust management technology based on behaviors is a new research hotspot in recent years, which is the effective key solution to security issues of a new distributed application mode. Therefore, it has important practical significance and wide application prospects to systematically and thoroughly carry out the research of the quantification mechanism of the dynamic trust relationship and to design the dynamic trust evaluation models for different distributed environments.Through the in-depth studies of the trustworthy network and cloud security that are new research hotspots of the network security, we establish rational trust models referring to different needs of trustworthy network and cloud computing environment and gives detailed and reasonable formal descriptions and inferential systems. For the trust model of trustworthy network, we consider multi-dimensional affecting factors to evaluate the complexity of trust relationship, introduce a sliding window to achieve dynamic trust assessment, build a Direct Trust Tree to store interactive relationship between nodes and adopt improved traceback algorithm to search the optimal recommending route based on the idea of greedy algorithm. For the trust model for cloud computing, a new method for users'trust evaluation based on fuzzy AHP has been presented and multiple detection engines have been used to conduct a comprehensive inspection of suspicious files and integrated decisions have been made. Experimental results show that our modles can reduce the additional overhead of monitoring and prevention caused by the mistrust, decrease the incidence of malicious attacks, enhance the whole network security and provide great security defense for uses.