Detection of Distributed Denial of Service Based on Fractal Feature of Network Traffics
PU Ru-feng · Xihua Daxue xuebao. Zhexue shehui kexue ban · 2009
Aiming at the difficulties in traditional procedure,this paper presents two new methods of network traffic abnormity detection for DDoS.From an investigation on the fractal parameters of network traffics,Hurst and it's time variable function Hurst(t) and Holder and it's time variable function Holder(t),the changes in Self-Similarity and Multi-fractal brought by traffic abnormity are analyzed.The detected abnormal changes in Hurst(t) and Holder(t) are used to determine network traffic abnormity.The investigation indicates that the statistical analysis of net-work traffic will benefit the detection and protection against the attack launched by DDoS.