网络攻击风险检测模型仿真研究  

Network Attack Detection Risk Model Simulation Research

作　　者：林德强[1] 

机构地区：[1]北京联合大学应用文理学院,北京100191

出　　处：《计算机仿真》2016年第1期305-308,共4页Computer Simulation

摘　　要：对网络攻击的伪装行为进行准确的检测,是网络安全的重要保证。传统的检测方法只能以模式识别的观点对已经出现过的网络攻击进行识别,对伪装出新特征的攻击,只要不在其建立的特征库中,就无法进行检测,缺陷较大。提出基于免疫算法的网络攻击风险的检测方法。根据免疫原理建立一种新的网络攻击检测器模型,获得可变化的抗体(检测器)的进化方向和病毒伪装的变化趋势;对检测到的疑似网络攻击行为,利用算法中的血亲原理进行分类,利用疫苗接种的机制对网络攻击的风险进行检测,提高了高伪装攻击检测的准确性。仿真结果表明,改进算法能够提高网络攻击风险检测的准确率,降低了虚警率。Camouflage of network attack behavior accurately detection, is the important guarantee of network secu- rity. Traditional detection methods can only from the viewpoint of pattern recognition has been in the cyber attacks, which can identify new characteristics of camouflage, as long as is not set up the characteristics of the library, can＇t test, defect is bigger. Put forward the risk of network attack detection method based on immune algorithm. According to the principle of immune to establish a new model of network attack detectors, access to change evolution direction of antibody （detector） and the change tendency of the virus in disguise ; To detect suspected cyber attacks, using the algorithm of the classified principle of consanguinity, using the mechanism of vaccination to detect the risk of network attack, improve the accuracy of the high disguised attack detection. The simulation experimental resuhs show that the improved algorithm can improve the accuracy of network attack risk detection, reduce the false alarm rate.

关 键 词：网络攻击 风险 检测模型 

分 类 号：TP393.4[自动化与计算机技术—计算机应用技术]