On Mon, Jun 06, 2016 at 07:19:31AM -0400,
 Tim Wicinski <[email protected]> wrote 
 a message of 79 lines which said:

> We started the discussion a few meetings back that we are planning
> on recharting to address the resolver-to-authority session.  We
> (warren and myself) wanted to wait until we've started seeing
> deployment and have gotten some operational data sets that would
> give every the warm fuzzy feelings needed to move along.

In the mean time, if someone is brave enough to write an individual
I-D describing the resolver-to-authority usage of TLS. Distribution of
the keys is of course the biggest difference with RFC 7858. The
security/authentication model has to be completely different. I see
several solutions:

* encoding the key in the auth. server name (as in DNScurve)
* publishing keys in the DNS, secured with DNSSEC (as in DANE), which
  raises an interesting bootstrap problem,
* not checking the keys at all, accepting anything,
* add here your own favorite solution.

Documenting these possible choices would certainly help and we don't
need rechartering to write/read an individual draft :-)

_______________________________________________
dns-privacy mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dns-privacy

Reply via email to