The Implementation Of Intrusion Detection System And The Application Of Neural Network

Based on the deep study on intrusion detection technology, an intrusion detection system building assistant system SIBAS is designed and implemented in this article. SIBAS can greatly reduce the labor for researchers to examine a new intrusion detection algorithm by providing a serial of high reusable modules and basic facilities. So the researchers can concentrate more on the key part of research.A rule-based network intrusion detection system is designed and implemented on the basis of SIBAS, which is named SIBAS-E. The successful developing of SIBAS-E shows that SIBAS is a successful design and on the same time, SIBAS-E is an expansion of SIBAS. Rule-based intrusion detection system which has the advantages of high precision and low mistakes is the main technology for implementation of commercial intrusion detection system. By the carefully designing of rule, SIBAS-E can easily adopt the rule base of the excellent open source NIDS SNORT and yet supply far more abundant expressing ability. There are about 930 rules in the rule base which makes SIBAS-E an useful system.Artificial neural network is a powerful tool. Some research has been carried on in the intrusion detection field. After a short review of the research work before, a new idea of applying NN to IDS is given. The idea is to build model of remote host's TCP/IP stack so as to detect general intrusion that violate the TCP/IP protocol integrity. A raw experimenting system is built based on SIBAS-E. And the result is inspiring.Based on the MIT Lincoln Laboratory intrusion detection system test set, a whole test plan for IDS is provided which gives foundation for evaluating of the intrusion detection method.