Network Boundary Security Situation Prediction Based On Time-varying Markov Model

The security situation faced by the network boundary is not optimistic while network security incidents occur frequently.If we do not understand the security situation of the network boundary,it will make the information easily stolen,damaged or attacked.Based on this,this paper carried out research of network boundary security situation prediction based on time-varying Markov model has important academic significance and application prospect.Firstly,the paper analyzes the overall situation of the network boundary security,and makes a research on the current situation of the commonly used security situation prediction technology and the Markov model.According to the security requirements of network boundary and network boundary,a security situation prediction framework is presented.The research will be carried out in two aspects,in combination with the secuurity situation trend prediction and the security alarm information prediction.The specific research contents are as follows:(1)According to the deficiency of the weighted Markov chain model in security situation prediction,taking into account the time-varying model,this paper introduced the"combination",forming a time-varying Weighted Markov chain model based on combination of states.The network boundary security situation prediction method based on the model is also presented.The DARPA2000 data is used as the simulation data to verify the model.Compared with the traditional time-varying weighted Markov chain model,the new presented model has higher real-time performance,robustness and accuracy of prediction.However,there are still some problems of stability.(2)Aiming at the disadvantage of poor stability,the Dempster theory is introduced.BPA is used to represent the state of uncertainty.The improved model is called a time-varying Weighted Dempster Markov chain model based on combination of states.Experiments show that the improved model can effectively overcome the lack of stability of the time-varying Weighted Markov chain model.(3)A security alarm information prediction method based on Hidden Markov model is given.The alarm data generated by the actual network boundary is collected as the experimental data.The validity of the method is verified and compared with the real security alarm information.Finally,all the work is summarized and analyzed,and the direction of future efforts are pointed out.