正态谐振网络病毒连续攻击下的重叠检测算法

发布时间：2019-03-17 14:21

【摘要】：新型的网络病毒的入侵和传播以正态谐振方式进行连续攻击,检测困难,需要对病毒攻击进行有效检测。传统的检测方法采用干扰攻击定位状态转移特征提取的层次性检测算法实现对病毒连续攻击的检测,算法没能自适应求解连续攻击下的特征选择参数,检测性能不好。提出一种基于窄带信号频谱偏移修正的正态谐振网络病毒连续攻击下的重叠检测算法。采用链重叠检测和向量化运算模拟正态谐振网络病毒的攻击特征和相关性质,实现对病毒的检测,通过Langevin方程描述,正态谐振网络的病毒连续攻击表现为一种双稳态非线性驱动多频共振模型,采用窄带信号频谱偏移修正方法,计算得到病毒节点连接概率以及个体的感染率,提高检测性能。仿真结果表明,该算法对病毒传播有较好的抑制性,提高了对网络病毒连续攻击下的检测性能和重叠检测能力。
[Abstract]:It is difficult to detect the intrusion and propagation of the new network virus in the normal resonance mode, so it is necessary to detect the virus attack effectively. The traditional detection method uses the hierarchical detection algorithm of state transition feature extraction of interference attack location to realize the detection of virus continuous attack. The algorithm can not adaptively solve the feature selection parameters under continuous attack, and the detection performance is not good. An overlap detection algorithm based on narrow-band signal spectrum offset correction for continuous virus attack in normal resonant networks is proposed. Chain overlap detection and vectorization are used to simulate the attack characteristics and related properties of the normal resonant network virus, and the detection of the virus is realized. It is described by the Langevin equation. The virus continuous attack in the normal resonant network is a bistable nonlinear driving multi-frequency resonance model. The narrowband signal frequency shift correction method is used to calculate the connection probability of virus nodes and the infection rate of individuals to improve the detection performance. The simulation results show that the algorithm has a good suppression on virus propagation and improves the detection performance and overlapping detection ability of network virus under continuous attack.
【作者单位】：河南中医学院网络信息中心;
【分类号】：TP393.08

【参考文献】