基于SDN的流量异常检测与预测算法

Based on SPE and EWMA algorithm, a multi-function traffic monitoring algorithm L-EWMA which can be applied to real SDN network scene is designed. The algorithm archives real-time detection of network anomaly and automatically judges according to the characteristics of abnormal traffic types(Alpha, DDoS and Node Loss). Compared with other SDN anomaly traffic detection algorithms, this system adds the functions of abnormal traffic location and prediction analysis. Experiments show that L-EWMA have high accuracies in analysis abnormal scenes, location and prediction flow change. Through this algorithm, the technicians can fully control the traffic situation of the network and make accurate judgments.