On 23/01/2020 22:07, Paul Vixie wrote:

> apex cname is a bad idea that can't be stopped. modern dns has a lot of 
> these, 
> for example ECS and what i once called "stupid DNS tricks"[1]. anything that 
> anybody wants to do is a possible topic for wide deployment and eventual 
> standardization, and system coherence be damned.
> 
> HTTPSSVC is the right answer for this. we should push _hard_ on that rather 
> than distracting our energies with trying to fix the DNSSEC problems related 
> to apex cname. to DNSSEC, a name is either canonical or not. if it's not it 
> will have a CNAME and only a CNAME -- and this is a good thing.

+lots!

I hoped my HTTP record draft might have been the solution, but even
though that's now dead I like to think that it helped catalyze the
browser folks towards HTTPSSVC.

CNAME was *never* the right answer for directing traffic for a domain to
a specific host, but it happened to work and was the only tool in that
toolbox at the time.  Attempting to extend it to the apex just makes
that worse.

Ray


_______________________________________________
dns-operations mailing list
[email protected]
https://lists.dns-oarc.net/mailman/listinfo/dns-operations

Reply via email to