Research On Attack Entropy Optimization Algorithm Based On SNAG Model

With the advent of the Internet of things era,network structure is more and more existing in the real world.At the same time,the challenge of attack in network structure is also changing.The degree of attack threat faced by the core node in the network structure changes with the attacker’s attack experience,and the reachability probability of the starting node may change with time.In the current context,how to effectively predict the most likely attack path in the network ahead of time has become a problem in the current Cyberspace Security.In this paper,aiming at the problem of eliminating redundant paths and real-time prediction of attack paths,firstly,a sequential network attack graph model is defined on the basis of predecessors to predict the whole attack process;secondly,in order to make the elimination of attack paths more reasonable,attack entropy is defined to measure the attacker’s own attack experience;finally,we use the connection tree algorithm in the FullBNT library of MATLAB to calculate the confidence of each node.Through experiments,it is proved that the model can predict the confidence at any moment in the attack process;by comparing the attack entropy parameters with or without,The value of risk cost is smaller.The main contents and innovations of this paper are as follows:1.In order to predict the best attack path at each time in the attack process,this paper defines the starting probability function,and combines the time and the network attack graph at every time,and defines the sequential network attack graph model.This model can simulate the whole process of network attack,and then can predict the best attack path in advance.2.In order to measure the impact of the attacker’s attack experience at the risk cost,this paper defines the attack entropy by using the number of nodes occupied by the attacker in the process of multiple exploratory attacks.The risk cost calculated by the attack entropy and the attack experience gained by the attacker in the attack process are considered,so the risk cost is more reasonable.3.In order to quickly and accurately infer the confidence degree of each node in the network attack graph,this paper applies the connection tree algorithm in the precise inference to calculate the confidence degree of each node in the network attack graph.The connection tree algorithm has less Monte Carlo simulation process than the previous likelihood weighted method,so it is faster in the confidence calculation of attack graph nodes.