分布式网络异常攻击检测模型仿真分析

针对传统的异常攻击检测方法主要以异常攻击行为规则与网络数据隶属度大小进行判别,只能针对已知异常攻击进行检测,对新型异常攻击,检测算法率低,计算数据量大的问题。提出一种新的分布式网络异常攻击检测方式,通过对分布式网络内数据进行迭代聚类将正常和异常数据进行分类,建立矩阵映射模型进行数据矩阵对比,初步对异常攻击数据进行判断。在矩阵中建立粒子密度函数,通过粒子密度变化计算其异常攻击概率,最后对其数据进行加权和波滤确定数据异常攻击特征,建立攻击检测模型。仿真实验表明,优化的分布式网络异常攻击检测模型提高了异常数据攻击检测的自适应性,在网络信号受到攻击信号干扰情况下,仍然能够准确检测出带有攻击特征的小网络异常数据。有效提高了分布式网络的检测正确率,加快了检测速度和稳定性。

istributed network anomaly detection model in the simulation analysis

In view of the traditional attack detection method mainly anomaly attack behavior rules and network data membership size discrimination, can only detect abnormal for known attacks, to new anomaly attack, low rate of detection algorithm, calculation of large quantities of data. Put forward a new way of distributed network anomaly detection, through the study of the iterative clustering of data in a distributed network classify the normal and abnormal data, data matrix comparison matrix mapping model, preliminary judgment on abnormal attack data. Establish the particle density function in the matrix, through calculating the particle density change abnormal attack probability, finally the data on the weighted and wave filter to determine abnormal data attack characteristics, attack detection model is established. Simulation experiments show that the optimization of distributed network anomaly detection model improves the adaptability of abnormal data attack detection, attack in the network signal interference circumstance, still can accurately detect the small network attack characteristics of abnormal data. Effectively improve the detection accuracy of the distributed network, accelerate the testing speed and stability.