The Design Of Network Intrusion Detection System (NIDS) And Research Of Interaction With Firewall And NIDS

With the rapid development of Internet, Internet has gone deep into each field inour daily life;meanwhile, the connectivity sharing and distributing of the Internethave increased the fragility and possibility of the network security. Therefore,intrusion detection has received great concern as a kind of effective measuring.Intrusion Detection System (IDS) can take measures in real time to computer systemand network and then find the intruder who breaks into the system or network. Thisdissertation firstly introduces intrusion detection technology and intrusion detectionsystem. And then presents problems intrusion detection system faced and completesthe procession of implementation of IDS. Artificial neural network (ANN) has special advantages in the intrusiondetection system with the ability of abstract conclusion, the self-adaptive ability andthe internal parallel computation. Using neural network, data mining can extractimplicit, unknown, non-ordinary and potential application valuable information andpattern using neural network to help intrusion detection system to carry out decisionanalysis. This dissertation presents an application model using recurrent network inintrusion detection system based on principal component analysis and gives itsexperimental result. Intrusion detection system has the function of finding intrusion and cutting offconnection, but its emphases is more put on the discernment of intrusion behavior.Security policy of the whole network should be configured by firewall. Intrusiondetection system should interact with firewall, dynamically changing the policy offirewall, and completely cutting off the behavior of intrusion from intrusion sourceusing firewall. This dissertation introduces a design schema which combines intrusiondetection system and firewall to work in coordination to achieve better networksecurity result. Finally, this dissertation summarizes job of this text. And we estimate the futuredevelopment of our research.