Research On The Intrusion Detection Technology In The Application Of The Computer Monitoring Systme Network

With the acceleration of construction of railway information, more and more networked computers join with the supervision system, the original static network security system which is the core of firewall can not satisfy the current network security protection, in order to concern network intrusion and protective network security actively, accelerating the development in the supervision network safe protection is necessary. Meanwhile, as the main communication protocols of the supervision system, TCP/IP protocol is simplicity and openness, therefore the use of TCP/IP protocol of the vulnerability of network attacks is increasing, in order to protective network security preferably, design a good performance of intrusion detection system for TCP/IP protocol has important practical significance.By using the neural network's advantages of learning ability, good fault-tolerance, adaptability and real-time performance, combining the intrusion detection technology, the paper designs a network intrusion detection system based on BP neural network.First of all, test object of intrusion detection supervision system network in packets. The data into the neural network training and test is the normalized vector, the data must be preprocessed. The pre-treatment process includes protocol analysis, packet feature extraction and normalized. In the process of protocol analysis, for the network characteristics of supervision system, analysis TCP/IP protocol respectively; In respect of normalization, mainly adopts linear function conversion normalization.Then, feature extraction is for the data after pretreatment. This paper selects the KDD CUP99 which is structured by the MIT in 1999 as data sources. Extracting 41 dimensions characteristic and normalized, after that the data will be learned in BP network, extracting intrusion detection rules after trained and tested neural network, build rule storeroom, based on the matching results of rules to detect and analyze.Once again, the paper designs the BP network intrusion analysis module, aiming at the characteristics of supervision system network, mainly selects BP network structure, improves algorithm and sets initial parameters, and selects invasion characteristics of data samples for network training.On the basis of theoretical analysis, using Matlab programming language realizes corresponding algorithm.From the test results, intrusion detection system based on BP neural network can solve the network safe protection problems in supervision system and reduce the problem of the intrusion detection system the high rate of error and omit. For some unknown attacks, neural network can also recognize through self study.