Slashdot Mirror
Spoiled Onions: Exposing Malicious Tor Exit Relays
An anonymous reader points out this recently published study (PDF) on detecting malicious (or at least suspicious) Tor exit relays. From their conclusions: "After developing a scanner, we closely monitored all ~1000 exit relays over a period of four months. Wed discovered 25 relays which were either outright malicious or simply misconfigured. Interestingly, the majority of the attacks were coordinated instead of being isolated actions of independent individuals. Our results further suggest that the attackers made an active effort to remain under the radar and delay detection."
One of the authors, Philipp Winter, wrote a followup blog post to help clarify what the paper's findings mean for Tor users, including this clarification: "First, it's important to understand that 25 relays in four months isn't a lot. It is ultimately a very small fraction of the Tor network. Also, it doesn't mean that 25 out of 1,000 relays are malicious or misconfigured (we weren't very clear on that in the paper). We have yet to calculate the churn rate of exit relays which is the rate at which relays join and leave the network. 1,000 is really just the approximate number of exit relays at any given point in time. So the actual number of exit relays we ended up testing in four months is certainly higher than that. As a user, that means that you will not see many malicious relays 'in the wild."
"New information" being this isn't 25 of 1,000 nodes.
its 25 of some unknown number of nodes, of which 1,000 are active at any one time.
And as I tried to point out last tiime (and am greatful for the opportunity to reiterate)
exit nodes only account for 100Mbps of tors 3Gbps average traffic (most of the traffic being to hidden services which never go near an exit node)
So if anything this is testament to the security of tor.network.
I guess much of the fear comes from the silkroad take down, but that was foiled by the good old postal service and human error, not the technology itself.
25 out of 1000 relays were detectably suspicious. These are the script kiddies who set up an exit node in order to harvest credentials that can be used for fraud etc. Such nodes are easy to detect by verifying https certificates and/or transmitting false credentials over tor and checking if they are used later.
The really sinister exit nodes are not as easy to detect. Transmit false dissident names and check if the named people are imprisoned and tortured?
I am the main author of the referenced paper. We tested more than 1,000 exit relays but don't know the actual number yet. However, it can be determined based on Tor's historical relay descriptors. The reason that's important is because the naive statistic "25 in 1,000 were malicious" is wrong.