Study On The User Behabioral Trust In Multi-layer Distributed Network System

Trusted network is an extension and application of the trusted computing in network. With the rapid development of the e-commerce and e-government activities, people are increasingly dependent on the network. It is significant research point that how to construct a safe, controllable and reliable network. The network user behavioral trust is one of the most important content in trusted network to be researched. Compared with the trust of user's identity, the user behavioral trust not only can be controlled more specifically, but also a dynamic trust form. So the user behavioral trust is key factor for constructing the trusted network. An assessment algorithm of the network user behavioral trust and a mechanism to monitor and manage user behavioral evidence are proposed in this paper. And then, a multi-layer distributed network system is constructed to verify the research results above.The transitional network user behavioral trust is a fuzzy concept. So in this paper, an assessment method which based on fuzzy decision analysis for network user behavior is proposed. By calculating the user behavioral membership degree of trust (the trusted membership degree), and grading the user behavioral trust according to the trusted membership degree, thereby the assessment results of the user behavior are achieved. Advantages of the method are that it is fit for the ambiguity of network user behavior, has lower computational comlexity, and can be easily applied to application in the real network. The assessment result can be used to calculate routing and adjust user access permission.User behavior is characterized by the specific evidence of the user behavior. User behavioral evidence is basal value to quantitatively assess the user behavior. Therefore, in this paper a mechanism is proposed to monitor and manage user behavioral evidence. The evidence monitoring mechanism includes network availability monitoring, network performance monitoring and security event notification from central server. All evidence obtained from monitoring is quantified as "excellent" membership degree. Then in order to achieve the management of user behavior, the "excellent" membership degree is weighted with the historical record membership degrees. This mechanism can effectively obtain evidence of the network user behavior and calculate the "excellent" membership degree of the user behavioral evidence.Based on these results, the Python programming language is used to construct and implement a multi-layer distributed network system which is depended on the assessment of the user behavior. The proposed system uses the research results mentioned above to distribute tasks and permissions. Experimental results show that the presented system can effectively monitor and manage user behavioral evidence, and achieve accurate assessment of the user behavioral trust.