| Internet communication security and privacy protection have received extensive attention,and a large number of anonymous network technologies have been proposed.Among them,Tor is one of the most widely used low-latency anonymous communication systems.However,while Tor protects user privacy information,it is also used by criminals.Construct all kinds of illegal activities.If the attacker’s attack behavior can be detected and intercepted at the Tor exit node,it can not only reduce the waste of Tor network bandwidth,but also enhance the enthusiasm of users to volunteer to become Tor node volunteers,and maintain a harmonious and safe network environment.Therefore,it is of great significance to detect the attack behavior of the Tor anonymous network.However,the network attack behavior detection of Tor is faced with the problems of inaccurate selection of traffic characteristics and unbalanced network traffic,which need to be solved urgently.This paper conducts research on this,and the main work is divided into the following three aspects:Firstly,this paper analyzes the anonymous communication system and the principle of Tor anonymous communication,summarizes several common types of network attack behaviors,investigates the key technologies of network attack behavior detection and common machine learning algorithms in this field.Then,aiming at the problem of inaccurate traffic feature selection,this paper proposes a network attack behavior detection algorithm THFS based on tree model hybrid feature selection.The THFS algorithm is based on the tree model method for feature selection,and combines the mutual information and correlation coefficient feature selection methods to extract effective features,reduce feature redundancy,and reduce complexity.Secondly,for the problem of unbalanced network traffic,the K-Means method is used to sample the balanced traffic distribution,and combined with the soft voting ensemble method,multiple classifiers are combined for classification to improve the accuracy of attack behavior detection.Subsequently,the effectiveness of the THFS algorithm in the detection of known attack behaviors is verified by experiments.Finally,this paper designs and implements Tor-THFS,a Tor anonymous network attack behavior detection system based on THFS algorithm.The Tor exit node service is analyzed,and on this basis,the requirements analysis and overall design of the Tor-THFS system are carried out,and the overall architecture scheme and functional module design of the system are proposed.The key function model of the system is tested to verify the integrity of the system function. |
PDF Full Text Request