带可信度评估的连续小波分布式拒绝服务攻击检测算法

DDoS Attacks Detecting Algorithm Based on Continuous Wavelet Transforms with Reliability Evaluation

针对传统方法难以实时、有效检测分布式拒绝服务(DDoS)攻击的问题,提出了一种带可信度评估的连续小波DDoS攻击检测算法.首先用不间断的连续小波变换对流量信号进行同步分析,通过发现平台突发信号来实时检测DDoS攻击,然后用报警可信度评估算法对经连续小波变换的检测结果进行二次处理,以消除单点突发信号和网络流量噪声带来的影响.经离散小波变换法、N点平均法以及梯度法的实验对比表明,所提算法对流量数据中的平台突发信号的检测效果比较好.

As traditional methods cannot effectively detect DDoS attacks in time, a DDoS attack detecting algorithm based on continuous wavelet transforms with reliability evaluation is proposed to detect the DDoS attacks in real-time. In the algorithm, continuous wavelet transforms are used to analyze the traffic data uninterruptedly and to detect the short-flat-burst or the long-flat-burst, which always represent DDoS attacks, and then an algorithm evaluates the reliability of alert, and to reduce the inaccurate alerts caused by single-point-burst and flow noise. Experiments show that the detection algorithm is more effective in detecting DDoS attacks than the discrete wavelet transform method,N-point-average method and gradient method.