As DoQ deployment increases, DNS becomes less constrained by the
traditional limitations of DNS over UDP. Combined with DPLPMTUD, that
may gradually reduce the pressure to optimize DNSSEC solely around
response size, while TCP remains a perfectly viable fallback.
—Carlos
On 07/07/2026 09:32, Joe Abley wrote:
On 7 Jul 2026, at 07:02, Ondřej Surý<[email protected]> wrote:
I believe it is dangerous to say "it is ok to deploy PQC algorithms with
large signatures" because we have these optimizations. The fact is
that we either need a PQC algorithm that will be size-suitable for DNS
or we will have to brace for the impact of completely switching to TCP.
I think this idea (Wes' idea) is interesting but I think we are again just
moving deckchairs. We could keep bolting on new bracing to prolong the life of
UDP transport, or we could just concentrate our efforts on the inevitable
widespread option of something better.
We have secure, stream-based transport defined and in use between stub and
resolver. It doesn't take too much imagination to see those transport protocols
being used in other parts of the resolution graph, with port 53 relegated to
legacy support for a long tail of clients that will probably never be able to
consume a PQC signature.
I don't think there is any chance of DNSSEC meeting any of the PQC deadlines we
know about. However, PQC is already deployed and in use in TLS. The encrypted
transport protocols already defined and in use in the DNS use TLS.
We should concentrate our energy on connecting these dots.
Joe
_______________________________________________
DNSOP mailing list [email protected]
To unsubscribe send an email [email protected]
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]
