Research On Fingerprint Attack And Defense Based On SSH Anonymous Website

Anonymous systems can protect the privacy of their users by encrypting data.Due to the elusive of anonymous systems,fingerprint attack have become an important tool for cyber regulators to identify unlawful activity in anonymous networks.At the same time,fingerprint attack is a double-edged sword that is easy to be exploited by unscrupulous elements to steal the privacy of users from anonymous systems,and the fingerprint defense model has become an effective way to further protect the privacy of users.The existing attack model can't deal with the data sets with abnormal missing data and high-dimensional features,and the existing defensive model has the defects that the selection strategy of fuzzy goal and the fuzzy rules are simple which lead to that there is a risk of being cracked.This thesis is based on the analysis of the upstream traffic of SSH anonymous communication system,and proposes more efficient models of fingerprint attack and defense.In the aspect of fingerprint attack,this thesis extract a number of novel data features of high contribution value,in the upstream traffic detection,Random forest classification algorithm combined with high dimensional feature data to construct fingerprint attack model for the first time,which achieved 96.4%attack accuracy through the verification of the experiment.In the aspect of fingerprint defense,this thesis puts forward a new heuristic defense method RTS,which is introduced for the first time in the selection of fuzzy targets,and more complicated fuzzy rules are formulated.The accuracy of fingerprint attack accuracy after defense is reduced to 6.7%.And in order that the fingerprint defense method is more suitable for a real network environment,improve the K-means algorithm and introduced it into the fingerprint defense,which reduces the bandwidth consumption of more than 16.8%while maintaining a certain effect of fingerprint defense.