On 20.5.2014 13:52, Chris Thompson wrote:
On May 20 2014, Mark Andrews wrote:
I've updated draft-andrews-http-srv-02.
http://tools.ietf.org/html/draft-andrews-http-srv-02
Wouldn't it be desirable to say something about https URIs as well as
http ones? It would seem that we will need an _https._tcp.[name] SRV
RRSet as well as the _http._srv.[name] one. (The idea of https overriding
the port number(s) in the _http._srv.[name] records with 443 seems
too horrible to contemplate.)
Hmm, would it be too weird to use
_http._srv.[name] CNAME _https._tcp.[name]
as 'HTTPS required' signalization?
(This is weird, I admit that. There will be troubles with DNS client libraries
not exposing CNAMEs etc... I just can't resist.)
--
Petr Spacek @ Red Hat
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
