Trust Model Research Based Risk Evaluation
|School||Nanjing Normal University|
|Course||Applied Computer Technology|
|Keywords||Risk assessment Trust model Punishment incentive Recommendation trust|
With the development of Internet, the network structure from the traditional centralized, static form gradually transformed into a distributed, open dynamic structure, the traditional security mechanisms can not meet the new demands of today's network development. To the distribution system of trusted computing challenges, trust the assessment and management of distributed systems has become a hot issue. This article will trust evaluation applications trust the dynamic nature of the relationship of trust and risk assessment of the risks to the network entity during the interaction, based on analyzing the existing trust model status and lack of in-depth study, for the credibility of spatial decision laid the foundation. This paper studies the following: 1) a new risk assessment model. In order to make the results more accurate trust evaluation, take full account of the factors that affect trust. In this paper, through the study of the relationship between trust and risk to determine risk factors is an important factor affecting the trust, proposed a system risk assessment model. Before the risk will be more simple and seen as a supplement of trust, a relatively simple calculation, no comprehensive consideration of the characteristics of the system itself and the characteristics of the physical behavior. In this paper, in-depth research and system entities threat event, the relationship between vulnerability and asset value of the security, put forward a framework for risk analysis, risk into the risk of confidentiality, integrity, risk and availability risk. According to this risk framework, this paper proposes using the Markov method to calculate the likelihood of risk, not only simple and accurate. Comprehensive of the three types of risk, this paper proposes using the the quantitative analytic hierarchy process to determine the risk weights, to solve the problem caused due to the security value of the different quantitative criteria. 2) a direct trust and recommendation trust calculation method. Trust assessment of this important addition to the risk factors involved in other critical issues, such as the calculation of direct trust and recommendation trust. evaluation. Calculation of recommendation trust for the selection of the recommended route, the three filter method, the recommended path is optimized improve interact satisfaction. Looking for credible Recommended entity is a prerequisite to obtain satisfactory service, this paper presents to distinguish the recommendation trust calculation method recommended quality service, recommended quality trust and quality of service recommendation trust trust. This method improves the previous studies due to confuse the two, mistaking services recommended by the high degree of trust and less capable entity as recommended by the lead interact lower satisfaction. In this paper, the physical similarity and frequently as a trusted computing factors, makes the assessment of the recommendation trust more accurate. 3) a risk of punishment incentives trust weighting function. Fusion punish the relationship between risk and trust, incentives, design risk belief function. The function of trust into account not only the trust and the level of risk, but also on the dynamic changes of the risk, the trust has a high sensitivity to risk, can effectively detect and defend against malicious behavior of the entity. Simulation results demonstrate the effectiveness and efficiency of the model.