一种分布式网络环境下基于挑战-响应模型的可信评估方法

A Trusted Evaluation Method Based on Challenge-Response Model in Distributed Network Environment

运用信任模型进行可信评估是解决分布式网络安全问题的重要手段。然而,目前大部分研究工作把研究重点放在如何收集更完整的信任证据,以及如何利用一些新手段如机器学习、区块链等评估节点信任值,很少对如何获取节点可靠的初始信任值进行研究。实际上,针对分布式网络提出的很多信任模型都依赖于历史信任证据,而初次对网络进行可信评估时并不具备相关历史信息。基于此,该文面向分布式网络环境的安全问题,提出了基于挑战-响应模型的可信评估方法。首先利用挑战-响应模型获取节点可靠的初始信任值,并利用此初始信任值对网络中的节点进行分簇,在簇内进行信任值计算和信任值更新,完成分布式网络环境下完整的可信评估流程。仿真结果表明,相较于统一设置初始信任值的方式,该文所提方法能对恶意节点、自私节点的信任值有较准确的预测,同时对恶意节点的检测率也更高。

Using trust models to conduct trust evaluation is an efficient way to solve the security problem in distributed networks. However, most of the researches focus on collecting trust evidence completely or using new methods such as machine learning, blockchain to conduct trust evaluation. Few of the researches focus on how to obtain reliable initial trust of network nodes. In fact, many trust models for the distributed network rely on historical trust evidence, but the historical information is unavailable for the first trust evaluation. To address this problem, a trust evaluation method based on challenge-response model is proposed. First, the challenge-response model is leveraged to obtain a reliable initial trust. Then, the trust is used for trust evaluation process, including clustering, trust calculation and trust update. Simulation results show that the proposed method has better performance than the unified initialization trust based method, in terms of the prediction accuracy for malicious nodes and selfish nodes, as well as the detection rate for malicious nodes.