Selective denial of service on guard nodes is a well-known attack, in which the malicious guard keeps dropping connections from Tor clients who are not in the set of his targeted victims.
Do Tor authority servers, or probing servers, or those clients who were dropped by the malicious guard have any mechanism to report it to the authority servers so that the malicious guard will be rejected in the next consensus vote?
Do Tor authority servers, or probing servers, or those clients who were dropped by the malicious guard have any mechanism to report it to the authority servers so that the malicious guard will be rejected in the next consensus vote?
Malicious relays should be reported to [email protected], along with the malicious activity, the fingerprint of the relay and steps to reproduce it.
However, if the denial of service were to be selective, like you suggest, then it would be difficult to independently verify the malicious activity.
If the relay is found to be acting maliciously then it will be blacklisted and removed from the consensus.
bad-relays mailing list (which is invite only) but I know that there are many scanners like exitmap (designed to find bad exits) and other scanners designed to find bad HSDirs. A lot of the cases of targeted denial of service would be part of finding guards for specific users or onion services...but making yourself the guard for that service or user is infeasible, so more commonly such attacks would be temporary and hard to reproduce.
Commented
Dec 29, 2017 at 5:47