Research On The Key Technology Of Trust Management And Model Evaluation In The Distributed Scene

Network security is the key scientific issue needed to be solved in the application of the distributed networks. Traditional strategies, such as PKI, access control, because of their central-dependence and poor scalability, are no longer suited for resolving security issues of distributed system, such as P2P, Grid, Pervasive computing, etc. Trust Management, serve as an "soft secure" mechanism, aiming at solving distributed security-related issues, are feasible methods for solving the security of trusted behavior.In this paper, the related issues of trust management and model evaluation in distributed application scenes are studied. The contents mainly include trust model based on the parameter modeling in distributed scene, parameter hierarchical evaluation of trust models, trust model evaluation algorithm based on trusted modeling process, trust mechanism and models evaluation for the Internet of Things (IoT). The main contributions of this thesis are as follows:(1) A distributed trust model based on parameter modeling is proposed. Current trust models differed in the definition and description of the trust mechanism. Each model focused on different functional features and had their advantages and disadvantages. It was necessary to explore the more general trust model to solve the security issues in the distributed network. This paper proposed a distributed trust model based on parameter modeling. Nine functional parameters were extracted after investigating current trust models. These parameters included flexibility, subjectivity, fuzziness, time decay property, transitivity, anti-attacks property, rewards&punishment property, sensitivity and scalability. Each parameter was modeled and integrated to form a comprehensive trust model. Analytical results indicate that the proposed trust model satisfies all the nine functional parameters and thus has fair universality. Experimental results show that the proposed trust model is reasonable and effective. Comparisons with previous algorithms indicate that the performance of the proposed trust model has been improved.(2) The evaluation of trust models. Varied trust models have been proposed for different network scenes, such as P2P, Grid, Ad Hoc, Pervasive computing and WSN. This paper proposes a mechanism for evaluating these trust models based on parameter modeling and analytic hierarchy process (AHP). The proposed method evaluates trust models from qualitative as well as quantitative aspect with the same nine functional parameters extracted from the trust relationship. The optimal selection of a trust model is realized mainly based on the overall evaluated values. Simulation analyses in various scenarios are performed. The results show that the proposed method is effective and reasonable.(3) Based on innovation points from (1), we propose a trust model evaluation algorithm based on trusted modeling process. The evaluated trust model was divided into five trusted processes, including trust establishing, trust modeling, trust computing, trust decision-making and trust transmitting, according to trust lifecycle. Then each process is analyzed for reliability, simultaneously the fuzzy theory and Bayesian method were used for assessment and fuse to achieve the comprehensive quantitative evaluated results. Finally, the proposed method was validated in the evaluation of distributed WSN trust models. The analysis and simulation’ results show that the proposed algorithm is feasible and effective. The evaluated result laid a theoretical foundation for the optimal selection of trust models in the practical application.(4) Research on the key technology of the trust management in the IoT scene. We establish a trust management framework based on a layered model. The IoT is decomposed into three layers:sensor layer, core layer and application layer. Each layer is controlled by a trust mechanism and fused into the overall decision-making framework. In addition, service requester’s policy is also added into the final decision-making. Finally, we use fuzzy set theory and a formal semantics-based language to perform the abovementioned layered trust mechanism, also we discuss the preliminary trust model realization for each layer. The result provides a general framework and significant reference for the development of sound trust models for IoT.(5) Research on the trust model for the Internet of things from two aspects. On one hand, we evaluate several traditional trust models for WSN in the context of IoT. The evaluated parameters are extracted from a concrete context, which is establishing and updating of trust relationship between remote nodes after the WSN being integrated into the IoT. The results of evaluation and the simulation show that possible trust models are suit for IoT on condition that several restrictions are available. The evaluated method and results provide a reasonable reference for the establishment of trust model for IoT. On the other hand, we take the IoTs as a whole, and establish a quantitative trust model for the Internet of Things. Based on the previous layerd network model and distributed trust management structure, we perform trust modeling and trust calculating mechanism for each layer, and fuze to an overall trust model that is suitable for the Internet of Things. Finally, integrated with previous reasearches, we establish a secure solution that integration of trust management for the distributed IoT scenario.This paper is aiming at extracting generality from varied and various trust models, and contributing to standardization of trust and trust management from aspects of parameter modeling and trust model evaluation. Moreover, it is designed to reflect the security needs of the Internet of Things from two dimensions of communication identity security and trusted network node behavior, and proposes a set of trust related theories and methods suitable for the heterogeneous distributed IoTs. The results further provide theoretical support for the engineering application of the Internet of things.