MANET Invasion Examination System Research And Realization

Currently, wireless network technology is used in lots of different areas. comparedto a wired network, wireless network have some advantages such as the ease ofinstallation, the use of flexible and easy to extend and so on. However, behind theseadvantages, the wireless network is also facing a more serious security threat. MANET isa typical form of wireless network, which is a wireless network base station with arelatively traditional non-center structure communications network. In recent years, withthe Internet of Things technology, MANET network involved in the Internet has becomea trend. The MANET network intrusion detection system can effectively protectMANET.First, this article has carried on the comprehensive system summary to the wirelessnetwork invasion examination.The content including the topic background and thesignificance, infinite goes to prison to examine the system present situation and so on.Then, introduced the wireless network invasion examination system necessity, hasanalyzed the invasion examination architecture, has laid the foundation for the followingchapter.In view of the MANET network dynamic and open as well as node computationmemory resources finiteness, this article proposes based on the IDS proxy cooperationtype hierarchization invasion examination system. The system monitors the platform twoparts of constitutions by the IDS proxy and the invasion event. The IDS proxy has six bigfunction modules, respectively is: In the data collection module, the bunch cooperates theexamination module, the bunch cooperates in the examination module, the bunch invadesthe response, the bunch invades the response and the secure communication. Amongthem, in the bunch cooperates the examination is a hybird of BP and Elman neuralnetwork model, the model consist of real-time classification and event memory;Hierarchy reflected in the maximum connectivity degree algorithm by MANET nodes inthe network clustering, and the introduction of the cluster head node, IDS agent runningonThe cluster head node of the cluster is responsible for intrusion detection;Collaborative reflected in the different clusters of cooperation between the cluster headnode intrusion detection; System has two different detection process, namely, the cluster detection and cluster detection, cluster detected by the IDS agent from local routing tableand its received from outside the cluster and the cluster network traffic characteristics ofinterest to obtain the value of IDS, after pretreatment can be converted to a neuralnetwork model for identification of feature vector, then the neural network intrusiondetection. Inter-cluster testing is shared between the cluster head indecent intrusioninformation, and further to determine whether the node is a malicious node.we use NS2simulation of the performance of the system was evaluated by ROCcurve to determine the optimal value of system parameters. The results show excellentperformance of the system can be better applied to MANET network intrusion detection.