Intrusion Detection Through Network Performance Learning

With the development of the application of internet, corporations,goverments and universities all have build their own websites and provided some services such as E-Business,E-government,remote education and so on. But there is so much deleterious information on the Internet. Because of the shortcoming of the begin design of the Internet,there are so many security problems rising during the development of the Internet.If we do not resovle them,these security problems would take Internet into hell.So network security becomes the sticking point of Internet applications.In order to enhance the security of Internet,provide steady and safe services,a lot of security technologies and measures are used to protect the networks.Intrusion detection is one of those technologies that have done well in this field.First,the thesis analysis the Internet security situation and list the new challenges with the change of network situation.Then,the thesis introduce intrusion detection system, concept,functions,working principle,communication protocol,IDS product and evaluating standards.Importantly the thesis intruduce some intrusion detection methods that are being applied in the reach field now.And the thesis analysis the difficulties that intrusion detecton faces in the new network situation.This thesis provide the intrusion detection model through network performance learning.The thesis provide the structure figure and flow chart.And genetic algorithm applied in this model eliminate the high wrong alarm rate,the thesis provide coding policy,fitness function and object function.The thesis introduce a 3-tuple to describe the characteristic of the netvvork.The thesis provide the intrusion detection algorithm just to find intrusions.The thesis intruduce the program (99085) and the model that author have done in the program.In this work report here,the main contribution of the thesis is the network performance learning based intrusion detection model. This model done well in detecting the new attack and can work in the new network enviroment.