Greetings again. The current draft has a few issues that should be resolved 
before the WG decides whether or not to take on this work.

The following statement seems fundamental to creating the new mechanism in the 
draft:
   The ANY meta query was defined for debugging purposes mainly against
   resolvers.
However, there is nothing in RFC 1034 or 1035 that supports that statement. The 
following two sentences have the same problem:
   There have been widespread misunderstanding as to what
   the query is supposed to do and when it is approriate.  The query is
   intented for testing what records for a particular name a resolver
   has in its cache.
Nothing in 1034 or 1035 indicate "debugging", nor that the response for a query 
of QTYPE=ANY has anything to do with the contents of the cache.

The document later says:
   By default
   the implemenations SHOULD be restrict it to localhost via ACL.
...with no support for the reason for the "SHOULD". No one has shown damage 
from the current default of allowing ANY queries. Yes, they might reveal data 
that a particular server operator doesn't want to release, and it is fine for 
those operators to use an ACL, but this "SHOULD" applies to all operators 
without justification, at least for ANY or RRSIG.

Proposal: remove the unsupported justifications involving ANY, and make the 
default restrictions for ANY and RRSIG "MAY" level.

--Paul Hoffman
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to