> Can you clarify what you mean by "entire zone" here, especially in > the context of multi-signers? Do you mean signed by all signers? > Or provided by all name servers? > > Or is it actually about different subsets of RRs in a zone being > signed with different algorithms?
I think the requirement is that for an algorithm that has the ZSA bit set, validators always need to receive at least one signature for that algorithm with every (signed) RRset. That implies that every NS for the zone has to return those signatures but that also all currently cached RRsets have to meet that requirement. So I guess such a bit can only be set some time after an algorithm roll. > >Downsides: this will fail if a single zone is signed with two different PQC > >algorithms. In that case, the ZSA bit cannot be set. At the moment signing a > >zone with two different PQC algorithms doesn't seem likely. Who knows what > >the future will bring. > > Why are PQC algorithms special with respect to the ZSA bit? In > other words, why is there no failure for setting the ZSA bit(s) > for a zone signed with two different classical algorithms? In theory there can also be a failure for classical algorithms. It is not clear how likely that is. For simple downgrade protection, it would require the validator to strictly prefer classical algorithms over PQC. That seems unlikely. Another, slightly more likely case is a validator that prefers hybrid, a multi-signer setup where some signers sign with PQC+ECDSAP256SHA256 and others sign with PQC+ED25519. In that case, hybrid validation would not be possible because the the ZSA cannot be set for ECDSAP256SHA256 and ED25519. _______________________________________________ DNSOP mailing list -- [email protected] To unsubscribe send an email to [email protected]
