An Anonymous User Identifying Scheme In The Darknet

In recent years,more and more people begin to pay attention to personal privacy protection.In this situation,anonymous communication systems such as Tor have become increasingly popular.However,the anonymous communication system is easy to be used by malicious users.Especially,the darknet,formed by Tor anonymous communication system,is a typical example.Criminals rely on the anonymity of the darknet to engage in illegal activities,without worrying being caught.To identify users in the darknet is of great significance to maintain network security.Aiming at identifying anonymous users in the darknet,this thesis proposes an anonymous user identifying solution.Firstly,a new network flow watermarking method of interval grouping centroid-based is proposed.This method divides the network flow into fixed time intervals,considering two adjacent intervals as a group,and adjusting the size relationship of the two interval centroids in a group,in order to embed watermarking.Then based on the network watermarking method,the attacker uses the controlled hidden services to embed watermark into the data flow.At the same time,the attacker attempts to extract the watermark from the data flow received by the user,and checks the extracted watermark information to determine whether the user is accessing the dark web,thereby cracking the user’s anonymity.In addition,in order to efficiently and accurately distinguish the user’s Tor traffic from ordinary traffic,this thesis proposes a Tor traffic identification method based on machine learning.Firstly,according to Tor’s anonymity mechanism and flow characteristics,ten main features have been selected,then we improve the C4.5algorithm by simplifying the selection of segment threshold of continuous attributes and optimizing the time-consuming Logarithmic operation.The experiment results show that this method can effectively improve the accuracy of Tor flow identification.In this thesis,by analyzing the concealment and reliability of the watermarking scheme,we prove that it is suitable for the Tor anonymous network,and compare with the existing scheme in the simulation experiment,the result shows that the watermarking scheme is effective,and it is feasibility to identify user’s information in the dark web.