[saag] Re: on derivative work rights statements in emails to Security Area mailing lists

On 11/21/2025 10:07 AM, Nico Williams wrote:

> Speaking of which, I've not followed all your appeals, but regarding the
> non-hybrid PQ ciphersuites, I'm afraid that ship sailed the moment the
> IANA registry was made Specification Required.  I.e., in that particular
> case neither the WG chair, nor the AD, nor the IESG can provide relief.
> I share your concerns about that case FYI!  IMO we should absolutely not
> allow non-hybrid PQ for at least a few years out of an abundance of
> caution, but, again, I think it's too late and we can't really bring the
> horse back to the barn.

I think this boils down to an argument about power and authority of the 
IETF. Should the working groups act as gatekeeper and use their 
authority to "absolutely not allow" deployment of algorithms that they 
find inappropriate for the Internet? That would mean, of course, 
instructing IANA that only algorithms approved by Working Groups should 
be allocated code point. The problem with that is whether the IETF has 
enough moral authority to make such decisions stick. Sometimes it does 
-- see for example the rejection of various proposals to weaken 
encryption in TLS. If a substantial consistency wants to deploy 
something and the only thing standing in their way is a code point 
allocation experience shows that they will just squat on some random 
value, thus leading to a worst of both worlds. The IETF did not stop the 
deployment, and the squatted code point ends up having to be reserved to 
avoid collisions with existing deployments.

That's why I personally think that the IETF should concentrate on making 
sure that safe algorithms can be easily deployed, rather than attempting 
to use the IANA as a gatekeeper. In the particular dispute between 
hybrid and non-hybrid, that means publishing and endorsing a good hybrid 
solution.

By the way, the debate is not just about hybrid versus non hybrid. There 
is also a debate about how much deference the IETF should have to NIST. 
If we end up only recommending the technologies that NIST approved, then 
NIST becomes a single point of failure. Maybe we are OK with that 
because there are no credible alternative. Or maybe not.

-- Christian Huitema

_______________________________________________
saag mailing list -- [email protected]
To unsubscribe send an email to [email protected]