Re: [DNSOP] CloudFlare policy on ANY records changing

[Evan Hunt <[e-mail]>]

On Wed, Mar 11, 2015 at 08:05:31AM +0800, Paul Vixie wrote:
> if we're serious about redefining ANY as a meta-query, then answering
> with RCODE=0/ANCOUNT=0 is correct. (as it would be for RD=0 queries
> against an RA=1 server.)

I'm concerned that a NOERROR/NODATA for qtype=ANY, once cached, would be
identical to the cache representation of an empty nonterminal node, and
that all subsequent queries for any other qtype would then be answered
with the cached NODATA.

> but whatever we do, any new reaction to QTYPE=ANY has to ensure that the
> client gives up, and stops asking.

I liked the sound of REFUSED because it matches *XFR semantics, but if
you're right about how resolvers will react, then it's a bad idea.

I'm running on supposition here, and we need data. I wonder aloud if
CloudFlare or some other site would be open to trying a variety of
response semantics to find out which one pinches least.

-- 
Evan Hunt -- [e-mail]
Internet Systems Consortium, Inc.