Research On Web Recognition And Confusion Technology Based On Website Fingerprint

The traditional data encryption mechanism only hides the communication content,does not hide the traffic flow size and direction of metadata information,these metadata information seems to be harmless,but easily lead to traffic analysis(TA)attacks.With the continuous development of privacy protection technology,a variety of anonymous communication system(Tor is the most commonly used anonymous communication system)application born,it is both blurred the communication content,but also blur the metadata,to prevent a passive attacker Observe the source address or destination address of the communication.Research shows that even if the use of privacy protection technology to hide the contents of the packet,address,size,Website Fingerprint(WF)recognition is also possible.Network security threats include active attacks and passive attacks,and Website Fingerprinting recognition is a passive attack that allows local passive eavesdroppers to infer the user's Web browsing behavior through packet sequence information,which requires a small amount of computation and the risk of being detected very small,so it has brought a serious threat to the user privacy.Users who use anonymous communication systems do not want their web browsing behavior to leak to the eavesdropper,and need to defend against WF to steal user privacy.However,anonymous communication systems may also cause anonymous abuse,such as the use of anonymous communications system for criminal activities,such as foreign gambling,browsing pornographic information,the release of reactionary information,so it is necessary for these illegal activities in real-time monitoring.In the same kind of research,WF research was carried out under experimental conditions.They carried out a large number of conditional assumptions and did not take into account some of the user's web traffic behavior.In practice,these would happen.Based on the above problems,this paper presents a new type of Website Fingerprint recognition scheme SumSubsection(SSS),which can extract the characteristics of packet flow and give different weight,which is higher than the accuracy of the proposed Website Fingerprint recognition The WF recognition and Webpage Fingerprint recognition were compared and found that they are not the same concept.A successful WF recognition requires only a small amount of data to be trained.Recent studies on WF claim to have found a highly effective identification of Tor,but these studies assume some settings for the client,the attacker,and the Web.These assumptions do not necessarily apply to the actual situation,and this is the case for these multiple assumptions the study.Also found that the user's actual Web browsing behavior can further increase the success rate of WF recognition.A model is established for the cost of WF recognition.WF's real problem is not its uniqueness,but its connectivity,each visit to the site can be linked to the same fingerprint,based on this discovery,design and implementation of a novel website fingerprint confusion solution-randomized privacy modifier.Randomized privacy modifier to randomly send random time,random length packets,and randomly requested a false website as a noise flow,all aspects of the site fingerprint information was vague,it is better security,in addition,it is also a versatile confusion solution.