On 10/12/2021 23.13, Daniel Kahn Gillmor wrote:
- whether to hard-fail or not
I don't think it's much worth to bother with authentication if on-path active attackers could easily cause automatic downgrade anyway.
One problem with discovery in non-downgradable schemes without parent signals is when a resolver encounters a larger NS set. To make sure that none of the servers supports encryption, you'd need to do quite a large amount of queries (and also revalidate the NS). And you can't even continue deeper before finishing all of them. That's quite a problem; we've had DoS CVE across multiple vendors just for discovering huge amount of NS addresses in such malicious cases for each single client query.
Without the "more difficult" parent signals, this could be done by requiring to indicate encryption support in the NS name (_dot label or whatever). We need to validate the NS set anyway, and costs to implementing resolvers would decrease, especially DoS attempts like above, though the number of round-trips per referral would still be significantly higher than for resolvers without ADoT.
--Vladimir | knot-resolver.cz _______________________________________________ dns-privacy mailing list [email protected] https://www.ietf.org/mailman/listinfo/dns-privacy
