Ddos Attack Protection Technology Research

With rapid development of the Internet and its applications, network security has become a more and more important subject. Network attacks could damage networks, threat information security and cause huge loss to the users. Among all security affairs, Denial of Service (DoS) attack has become one kind of major events with characteristics such as extensive area, deep concealment, easy usage and high efficiency. DoS attack can greatly affect the effective service of the network and its host systems. Among various kinds of DOS, nowadays DDoS (Distributed Denial of Service) is seriously threatening the Internet, since it is more difficult to recognize and detect.Generaly,People often thinks that DDoS is becomes of the network system structure,for example network service public and network resource limited,and so on. All the factor make there isn't very good provention measure.At first, this papers will analyze various network attacking techniques and the network attacking will be growing ,and the network attacking wang to get thing. Second, we will concentrate on the study of DDoS. It is found that network flows has some self-similarity in long time scale. In addtion the basis of Tao Peng's research the server station would be visited by the same clients, Thereupon we will propose a method to fight against DDoS, which combines the network self-similarity and historical data. This paper will also analyze the LocalACC and PushBack algorithms, and use an improved method to trace DDoS attacking headstreams In the middle of network could throw away the attaching data.