Research On Dynamic Network Security Risk Assessment Method Based On Vulnerability Information

Computer network plays an important role in economic production and social life.However,in recent years,the number of computer vulnerabilities has been increasing rapidly,which makes network security face more and more severe challenges.It is of great significance to realize the quantitative evaluation of network risk in order to maintain network security more effectively.Therefore,this paper studies the risk assessment method of dynamic network security based on vulnerabilities.The main work is as follows.Firstly,the state attack graph is used to construct the evaluation model for the problem of network risk evaluation.It generates all attack paths in the network through the association vulnerability,which is helpful for the comprehensive risk assessment.In this paper,the state attack graph model is constructed by using network environment information and state attack graph generation algorithm.Secondly,for a large number of unknown vulnerabilities,the vulnerability category prediction model is established.Due to many categories of vulnerabilities and short description text,it is difficult to extract characteristic words.Therefore,the S-C feature extraction method is proposed in this paper.On the basis of defining the comprehensive function C which represents the importance of characteristic words,the information entropy S of words is used to weaken the importance of words with chaotic classification,and the words with large S-C value are selected to form the feature set.Combined with the bayesian algorithm AODE to establish the S-C-AODE category prediction model.Thirdly,in view of the previous risk assessment methods based on the state attack graph,the comprehensive factors such as attack revenue,attack capability and vulnerability release time have ignored the impact on the probability of attack events,as well as the prediction of risks.This paper proposes a model to dynamically evaluate network risks by quantifying the vulnerability hazards.This model uses the comprehensive factors to determine the state transition matrix,which makes the risk quantification more comprehensive and reasonable.By using the captured intrusionbehavior,the reachable probability of the state node is dynamically updated in real time,and the attack time matrix and defense time matrix are determined to realize the dynamic risk assessment and prediction.Finally,the S-C-AODE vulnerability category prediction model and dynamic network risk assessment method are tested and analyzed.