The Research Of The Intrusion Detection System Based On RBF Neural Networks

In recent years, the fast development of the computer network provides people tremendous facility, but at the same time, it also brings along the security problems. The firewall which has been the safest defending measure can not satisfy people's need. So the intrusion detection system (IDS) emerges as the times require. IDS effectively make for the deficiency of the firewall, it is a positive defense system which can protect against interior or exterior attack and mistaken operation, and before the intrusion invade our computer, it can timely response the intrusion and inform the administrator to take action.Because of the dependence of people to the network, more and more important information is spread through the network, the result is people have more request to the network security. But with the development of the defense system, the hackers do not stop create new attack measures, in contrast, it is more and more implicate, and the frequency and the number of the attack are also improved day by day, so our defense system have a good ability to detect the known and unknown intrusion. At present, the IDS lack of the self-adaptation and intelligence, and it does not satisfy the network of multiple attacks, so doing more research in this area is necessary.This paper points out the advantage of the RBFNN applied in the IDS and puts forward a model of IDS and the design thinking based on RBFNN on the basis of describing the network security,the intrusion detection technology,the neural network,RBFNN and the relative knowledge. This model effectively integrate the two detective technologies which is misuse detection and anomaly detection, unlike traditional system, we use five kinds of intrusions, two RBF neural network training module and three training mechanism. After experiment, we can see that this model have less training time and can detect the new intrusions in real time.In addition, on the course of research in this model, we find some problems, in the future, we should learn more and resolve these problems step by step.