Research On LDoS Defection Based On Harr Wavelet Analysis And Its Implementation

With the development of the network security, Some new methods and forms ofthe network attack technologies are invented, especially the technology of DOS (denialof service). This attack has been researched for several decades because it is every easyto implement and does great harm to the network security. There are some mature wayof the detection and defense of the traditional denial of service. In recent years, a type ofnew DOS attack, low rate denial of service attack (LDoS), which has a great effect thenetwork security and quality of service, appeared and has no widely accepted detectonmethods. So in this thesis the detection technology on LDoS is mainly researched.Firstly, the principle and classification of LDoS are summarized and analyzed, andthe comparison among the available detection methods is given. Then Harr waveletanalysis based LDoS detection methods and its implemtation are mainly researched forthe problem that LDoS has high attack efficiency and is hard to detect. In this methodwhether LDOS exists in a system is judged by Harr wavelet analysis according to theeffect of the attack on the network traffic and the number of reached packet per unittime. Some related problems of LDoS detection when the attack parameters arealterable are also analyzed in detail and some solving methods are presented. Finally,The implementing methods and each module design about the detection methods in realnetwork environment are given. The results of this experiment show that LDoS attack isable to be detected effectively by the methods.