基于多维指纹建模筛选机制的物联网反入侵算法研究  

Anti-intrusion Algorithm for Internet of Things Based on Multi-dimensional Fingerprint Modeling and Filtering Mechanism

作　　者：唐宝燕[1] Tang Baoyan(Weifang Vocational College of Engineering, Weifang, Shandong 262500, China)

机构地区：[1]潍坊工程职业学院

出　　处：《黑龙江工业学院学报（综合版）》2019年第10期53-58,共6页Journal of Heilongjiang University of Technology(Comprehensive Edition)

摘　　要：针对当前物联网反入侵算法存在入侵监测准确性较低,且难以适应动态入侵环境等不足,提出了一种基于多维指纹建模筛选机制的物联网反入侵算法。首先,基于数据挖掘思想,综合考虑攻击行为集合及其分布特征,采用匹配方式均衡节点缓存资源,分流攻击行为,提高网络在遭受入侵情况下的对攻击行为的命中,增强网络鲁棒性能;随后,针对入侵过程存在峰值特性,使用递归方式构建节点攻击行为模型,采取线性反序映射来获取攻击行为弹性系数并降低恶意节点入侵成功率,挖掘网络入侵行为的纳什均衡解,有效提高对恶意攻击行为的过滤效率。仿真实验表明:数据聚合-QOS联合检测算法(QoS-Aware Hybrid Data Aggregation Scheme,QA-HDA算法)、能量感知筛选监测算法(Energy-Aware Perception Algorithm,EAP算法)相比,在遭受入侵时,所提算法具有更高的临界性能及的入侵流量过滤能力。In order to overcome the shortcomings of the current anti-intrusion algorithm of the Internet of Things, such as low accuracy of intrusion monitoring and difficulty in adapting to dynamic intrusion environment, an anti-intrusion algorithm of the Internet of Things based on multi-dimensional fingerprint modeling and filtering mechanism is proposed. Firstly, based on the idea of data mining, considering the aggregation and distribution characteristics of aggression, the matching method is used to balance the caching resources of nodes, distribute the aggression behavior, improve the network's hitting and robustness in the case of intrusion. Then, in view of the peak characteristics of the intrusion process, the recursive method is used to construct the aggression behavior model of nodes. Linear reverse order mapping is used to obtain the elasticity coefficient of attack behavior and reduce the success rate of malicious node intrusion. The Nash equilibrium solution of network intrusion is mined to effectively improve the filtering efficiency of malicious attack. The simulation results show that, compared with data aggregation-QOS joint detection algorithm and energy-aware screening monitoring algorithm, the proposed algorithm has higher critical performance and intrusion flow filtering ability in the case of intrusion.

关 键 词：物联网 指纹建模 纳什均衡 反入侵 数据挖掘 线性反序映射 入侵流量过滤 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]