Research On Hidden Node Location Technology Of Tor Anonymous Communication Network

With the rapid development of technology,the internet has become ubiquitous in people’s daily lives.People’s awareness of privacy protection is increasing,and the demand for anonymous communication is also growing.The Tor network boasts low latency and high anonymity,making it the most popular anonymous communication network today.However,criminals can use anonymous communication systems to engage in a series of illegal activities,such as online fraud,sales of illegal goods,and dissemination of terrorist ideologies.This poses a great threat to social security and stability.This paper proposes a study on Tor network hidden node location technology from both the client and server perspectives to achieve precise location of criminals.Firstly,in response to the low accuracy of existing website fingerprint recognition models that rely on creating website fingerprint databases,we propose a website fingerprint recognition model based on CNN-GRU,which improves the process of extracting and classifying fingerprint features of user-accessed websites,thereby increasing the accuracy of location.The implementation of this solution ensures that accuracy and generalization are higher than several mainstream location technologies based on CNN and LSTM models.Secondly,in response to the low accuracy of existing flow feature recognition models that rely on creating flow feature databases to identify user identity information,we propose a flow correlation model based on CNN-ECA,which improves the process of associating flow features,thereby increasing location accuracy.The implementation of this solution ensures that the learning ability and ease of use are higher than several mainstream location technologies such as RAPTOR and PEARSON,under the premise of locating Tor network client nodes.