On Sat, Oct 24, 2026 at 11:54:54PM -0700, Wei Chuang wrote:
> I think the way to bridge this is to use the
> multiple signature support in DKIM2 and support switching between AND/OR
> evaluation of the multiple signatures as needed.
There's no need for any complexity of that sort. This is DKIM signing a
message that is in flight for at most a few days. It is not firmware
signing for a device that might last a few decades. Any single valid
DKIM2 signature supported by the relying party should be sufficient.
Unsupported signatures can be ignored (equivalent to missing).
If ML-DSA-44 is broken, it can be replaced fairly quickly, by a
classical algorithm if there are as yet no CrQCs that can
cost-effectively attack a ML-DSA-44 DKIM signatures. Otherwise,
by whatever PQ replacement algorithm is waiting in the wings.
If there's no viable PQ-replacement, in that hapless hypothetical
hybrids hardly help.
--
Viktor. 🇺🇦 Слава Україні!
_______________________________________________
Ietf-dkim mailing list -- [email protected]
To unsubscribe send an email to [email protected]