On 06-11-18 12:39, Ray Bellis wrote:
On 06/11/2018 17:58, Matthijs Mekking wrote:
That's the crux: A solution that depends on upgrading the resolvers is
considered not a (fast enough) deployable solution.
The HTTP record does not depend on resolvers being upgraded. If the
browser vendors implement the client side, it's not required.
But DNS providers like to solve the CNAME-at-the-apex problem within the
DNS protocol.
Once they do fully implement it by serving the A and AAAA records from
cache, then it'll be fast, too.
That's why I like ANAME: It allows you to do CNAME-at-the-APEX
processing without requiring resolvers to be updated, however
resolvers can implement ANAME to optimize the behavior.
Also the ANAME in its current form does not require (but also does not
prevent) the resolution to take place inside the name server, it can
be a simple script that is part of your zone provisioning.
I think Tony Finch was suggesting that you could also do that with "HTTP".
Okay, I missed that. If HTTP can do that too, than the approach is very
similar to ANAME except for the name. Why have both then? Also the name
HTTP suggests the record is only applicable to the web.
Matthijs
Ray
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
