Design And Implementation Of Network Attack Simulation System

In order to protect network security,with the support of machine learning and deep learning technology,the intrusion detection technology of intelligent network abnormal traffic was born.The intelligent model training was completed by using the network attack data set to optimize the detection ability of network attack.However,the accompanying problem is that the intelligent detection model lacks training data,resulting in poor classification effect.Based on the above problems,this thesis designs and implements a network attack simulation system to generate more real attack traffic.At the same time,it puts forward a method to generate traffic data using generation adversarial network technology and supplement the training data of intelligent detection model.The network attack simulation system uses the attack graph to guide the penetration attack route,uses multiple hosts to generate real attack data with the help of distributed technology,carries out penetration attack on the target network,and provides abnormal attack behavior input for the intrusion detection platform.The specific work contents are as follows:1)A simulation traffic generation method based on generation adversarial network technology is proposed.Through the adversarial training of generator and discriminator,the generator model that can fit the real traffic distribution is obtained,which provides training data for the detection platform,enhances the balance of training set and improves the accuracy of classification model.2)An attack path selection algorithm based on attack graph is designed.The attack graph is drawn through the vulnerability association relationship and the optimal attack path is obtained on the basis of the attack graph.3)The function of penetration attack is designed,which can scan the target host or target network,obtain host information,system information,service information and vulnerability information,launch penetration attack on the target host based on the collected information,obtain ordinary user permissions,improve permissions,log in to the target system as an administrator,complete penetration attack and provide abnormal attack behavior for intrusion detection platform.4)The distributed attack is designed.Using Kafka framework,the computing power and memory resources of multiple attack nodes are reasonably utilized through the load scheduling algorithm.And let nodes cooperate to complete scanning and attack tasks.5)Integrating the above multiple functional modules,a network attack simulation system is designed,and the implementation of the system is completed based on front and rear end technology and penetration technology.The system is divided into three parts:client,server and attacker.The client provides a visual interface for users,provides an interface for users to use system services and display the returned results;The server provides attack simulation function,attack path generation function and data storage function,and uses Kafka framework to schedule the nodes at the attack end;The attacker completes the scanning and attack functions of penetration attack.6)Through the test of the system,the generation adversarial networks are implemented,and training data set is supplemented which is realized to improve the classification accuracy of the model.It realizes the goal of launching penetration attack on the target network,generating attack traffic and providing abnormal attack behavior data for the intrusion detection platform.