Providing comprehensive and error-free anonymity to Tor users is at the center of academic research and technical discussion. From the technical point of view, the design of Tor architecture might look like it can achieve this goal; however, many issues make the system susceptible to failure. Some are related to user mistakes, some are onion routing issues, and some indirect issues affect the system’s success rate. (Due to the nature of the mechanism, Tor-related attacks refer to success rates that could be achieved. Not every user can be de-anonymized every time, but some users might be de-anonymized at some point in time.)
Types of attacks are covered in the next part in detail.
Leaving aside all non-Tor-related issues, the main subject of the anonymity research comes from monitoring the data transmitted on Tor. Then, the degree of anonymity could be measured via different models such as probability, similarity, entropy, and evidence theory based on the analyzed data. Since all network flow is encrypted between Tor relays with the help of these models, it might be possible to correlate the traffic and disclose the real IP addresses of the users.
Collecting data for traffic analysis, mostly encrypted, is the crucial step and of the utmost importance. Previous studies have focused on analyzing the network, collecting URLs of HTTP traffic, and so forth. A Tor exit relay creates another possibility here because anyone can operate an exit relay. The relay transmits the internet packages in an unencrypted format to the destination if the client is using HTTP instead of HTTPS (see section 2.2.3). Some researchers focused on this possibility, and some also used DPI to take a closer look at the data transmitting over the exit relay.
Post your comment on this topic.