网络多维信息安全融合技术的研究与仿真

Research and Simulation of network multidimensional information security fusion technology

传统网络多维信息安全融合方法无法获取节点的推荐信任,导致对正常节点与恶意节点的区分精度偏低,信息融合抗干扰性能较差。为此研究新的基于隐含关联度的网络多维信息安全融合方法。依据服务不同属性评价获取节点直接信任,利用隐含灰色关联度得到节点推荐信任,融合直接信任与推荐信任获取节点综合信任值;采用节点综合信任值区分正常与恶意节点,去除恶意节点的异常多维信息,根据权值融合正常节点的多维信息,实现多维信息安全融合。仿真结果证明,记忆因子取值为0.5时,新方法获取的综合信任值精度最高,可精准区分正常与恶意节点;上述方法融合多维信息的精度高,信息融合的抗干扰性与可信度分别高达98.5%与99.1%,验证了所提方法安全性高。

Traditional network multi-dimensional information security fusion methods have many defects due to the lack of node recommendation trust, such as low discrimination accuracy between normal nodes and malicious nodes, poor anti-interference of information fusion. Based on this, a novel network multi-dimensional information security fusion method based on implicit relevance was designed in this paper. The evaluation of different attributes of services was investigated for obtaining the direct trust of nodes. The implicit grey correlation degree was introduced to obtain the node recommendation trust. Direct trust and recommendation trust were fused to obtain the comprehensive trust value of nodes. The comprehensive trust value of nodes was applied to distinguish normal and malicious nodes, deleting the abnormal multidimensional information of malicious nodes. The multi-dimensional information security fusion was completed via fusing the multi-dimensional information of normal nodes with weights. The simulation results show that this method has high comprehensive trust accuracy(0.5 memory factor value),excellent anti-interference(98.5%) and 99.1% reliability.