[Doh] DNS64 and DOH

[Erik Nygren <[e-mail]>]

In reading through the dns-over-https draft I'm wondering if we should have
some better coverage of DNS64 which is becoming widely deployed in some
mobile environments.

(See https://tools.ietf.org/html/draft-palet-v6ops-nat64-deployment-00 for
some discussions which is on the v6ops agenda for this week.)

Thinking some about this (and general IPv6-only deployments):

* The section mentioning using IP literals as DOH service end-points may
wish to point out the risk that this may have challenges due to mixtures of
IPv6-only, dual-stacked, and IPv4-only environments.

* It would be good to add a section on how applications using DOH may wish
to handle "A" lookups in DNS64 environments.  In particular, to do the
DNS64-synthesis and construction in the client when no AAAA record is
available and when a NAT64 prefix is available through a mechanism such as:
           RFC 7050 - Discovery of the IPv6 Prefix Used for IPv6 Address
Synthesis,
           RFC 7225 - Discovering NAT64 IPv6 Prefixes Using the Port
Control Protocol (PCP)
 Doing the synthesis in the client seems preferable over trying to get the
server to do synthesis with the proper prefix in most scenarios, plus in
cases where the client is using the results directly this will get around
some of the DNSSEC issues.

At least for mobile, not using DNS64 synthesis in the client will likely
result in this either breaking the client in NAT64 environments or causing
something lower in the stack to do the equivalent (a 464xlat CLAT on
Android or bump-in-the-API on iOS).

        Erik