Research Of The Ad-Hoc Networks Intelligent Intrusion Detection System

Because of the open quality of the Ad-Hoc network, various ways to attack the network, and the improvement of the intrusion ways, especially the networks for the government and military to transferring information, people begins to think much of the security of the Ad-Hoc networks. At present, the passive network security mechanism, such as firewall, can't get a good effect, at the same time, as a kind of active whole defensive technology, intrusion detection can make up for the disadvantage of the traditional technology. It can detect the system user that uses the system without the certain right and the hacker attacks the system using the disadvantages of the security, through monitoring the whole network and the state of the system, the action of the user and the using information of the system. It can also take some certain solutions. Now, the distributed intelligent intrusion detective system based on multi-agent and ANN technology becomes the main point of this area. Multi-agent technology becomes one of the best useful technology in the intrusion detective system because it has the autonomic, interactive, adaptive and whole feature. ANN gets more attention in this area and becomes the main way of the mathematic description of the artificial cognition because it can predict the unknown network. With the application of this way, intrusion detective system can solve the security problem of the network and at the same time, it can solve the problem that the whole QoS drops in the network because of the attack. In this paper, on the basis of the present intrusion detective works, a distributed intrusion detection based on BP-CT algorithm was constructed which oriented the distributed attack –denial of service. In this system, agent was built in every node of the network. and the intrusion action was detected by the cooperation mechanism . a platform was built using the multi-agent technology. The ANN network technology and the CT algorithm were the engine of the intrusion detection, which could predict the probability of the distributed attack and could finish the final defensive work of the detection. The technology of agent, ANN, CT algorithm, QoS problem were discussed as well in the paper. In this paper, the security threatens of the Ad-Hoc networks were first analyzed and some concept related to the QoS was introduced at the same time. It also analyzed the present way to solve the intrusion detection system and brought up the disadvantages of the solution in the whole way. So it gave the solution to the solve the whole security problem of the network. At last, the intrusion detection system based on BP-CT was brought up which was based on the multi-agent, ANN, CT algorithm, and the artificial experiment was simulated. It proved that the intrusion detection system based on BP-CT can make a better defense to the Dos attack and the Dos attack could be predicted efficiently.