A flooding-based DDoS detection algorithm based on non-linear preprocessing network traffic predicted method

This paper presents an algorithm based on NLPP(Non-linear Preprocessing Network Traffic Prediction) for detection of DDoS attack. The algorithm, which is brought up based on the network characteristics on DDoS attack, consists of a basic detection method and a NLPP method. Compared to the classic algorithms, VTP-based detection algorithm and detection algorithm based on wavelet analysis, it is proven that is a more efficient algorithm concerning detection results. The algorithm can achieve 100% of both detection rate and accuracy rate when parameters are porperly set. Moreover, detection delay is greatly shorten and also it can realize well real-time detection of DDoS attack after NLPP is introduced.