Application DDoS Attack Detection Method Research

With the development of Internet technology,it has been formed the eco-economic circle with Internet as the core,and derived from a variety of Internet products.Most of these Internet products are based on the web server as a carrier.But hackers may launch application-layer DDoS attacks on these web servers and making these Internet products can not be used normally.In recent years,there have been frequent reports of application-layer DDoS attacks.Most attackers have targeted large commercial websites,financial institutions websites and government agencies' websites,causing a large amount of economic losses.Application layer DDoS attacks have become one of the major threats to the Internet,but the current detection technology can not meet the needs.Therefore,the efficient detection with application layer DDoS attacks is very importance for improving network security.This paper regards web server as the research object to design the detection technology with application layer DDoS attack.In this paper,we propose a detection model that combines quantum particle swarm optimization with logistic regression.Quantum particle swarm optimization algorithm is a kind of optimization.Its characteristic is to be able to achieve convergence in the global context and logistic regression is a classic dichotomous machine learning algorithm.The logistic regression model is trained by known samples,and predicted the test samples finally.We extract the octet feature in the user's access behavior as the logical regression model training samples or prediction samples,finally predict a certain type of user: attack or normal user.Finally proved by experiment,the Web DDoS detection model proposed in this paper can finally achieve a higher detection rate and a lower false detection rate.At the same time,it has a lower training time for the model and has higher detection performance.