Research On Dynamic Network Attack Behavior Prediction Method Based On Bayesian

With the development of Internet,Internet of Things and other network systems,the potential risks of potential networks are also increasingly exposed.Faced with an endless stream of cyber attacks,how the network system can run efficiently and smoothly has become the main subject of current research.Most of the existing network security metrics only consider the attributes of the vulnerability itself,but ignore the correlation between the vulnerabilities,and it is difficult to adapt to the quantitative requirements in complex dynamic networks.In addition,the complexity and uncertainty of cyber attack events are accurate and effective.Predicting defensive attack events and establishing a comprehensive real-time monitoring system poses difficulties.In order to solve the above problems,this paper conducts in-depth research on the prediction of dynamic network attack behavior based on Bayesian model.Firstly,based on the correlation and communication mechanism between vulnerabilities in the network environment,a new vulnerability risk measurement method is proposed.This method uses the causality of the availability and impact indicators in the CVSS basic indicator group according to the specific attack graph.Dynamically correct the availability index of the vulnerability of the child node,fully consider the inherent attributes of the vulnerability node and the environmental factors;Secondly,build a Bayesian two-layer network security analysis model based on Bayesian network,the model applies a new The quantification result of the vulnerability risk measurement method adopts the idea of Bayesian network and hierarchical structure.By dividing the probability attack graph into smaller-scale sub-networks,the probability value of the attack behavior in the Bayesian attack graph is effectively improved.Then,based on the Bayesian two-layer network structure model constructed in this paper,a new dynamic attack behavior prediction analysis algorithm is proposed.Based on the model,the Bayesian inference method is used to dynamically update the network attack.Confidence in behavior,prediction and analysis of attack behavior,full consideration of attack events on network systems The influence of the confidence of the attribute node in the middle;Finally,through experimental verification,the proposed method can effectively and accurately measure the vulnerability node and can effectively predict and analyze the network attack event.In summary,this paper proposes a new vulnerability risk measurement method based on the correlation between vulnerabilities in the network system,and applies the quantified results to the Bayesian two-layer network security analysis model constructed in this paper.A reasonable dynamic attack behavior predictive analysis algorithm can solve the problems existing in the existing network attack behavior analysis and prediction methods,and has good availability and high efficiency in a complex network environment.