一种基于小波求解的DDoS攻击检测模型

摘要

针对传统方法在检测DDoS攻击时的检测率和误检率的缺陷,提出了一种基于小波求解的DDoS检测模型.首先通过实时监控网络的数据流量,形成实时流量序列;然后根据得到的单位时间内的流量序列长度,动态更新分解尺度,对网络流量序列的长相关性的特征值hurst指数进行实时监控,以此来检测DDoS攻击.仿真实验证明,提出的检测模型能实时有效地检测到DDoS攻击的发生,检测率和误检率都较好,耗时较短.%Aiming at the bad detection rate and misdetection rate in DDoS detection using traditional methods, this paper proposed a DDoS attack detection model based on wavelet.Firstly, the monitor established the traffic serial according to inspect the network, then updated the analysis scale based on the length of the traffic serial, finally estimated the hurst parameter which was the main character of the long range dependence.So it could judge the server being attacked or not to check the change of the hurst parameter.Simulation results reveal that the proposed approach has a better detection rate and lower misdetection rate, and consumes less time.