short circuiting reverse lookups (resend)

[Aidan Williams <aidan.williams@motorola.com>]

itojun@iijlab.net wrote:
 >>Sure -- but to keep the load off the root, we need to be *very* sure
 >>that sites do pretend to be authoritative for them.
 >
 > 	will it help if we ship c.e.f.ip6.int/arpa zone files with BIND,
 > 	just like 1.0.0.127.in-addr.arpa?
 >

Yes, potentially.  Further, I would like to see a recommendation
that authoritative RFC1918 reverse maps are a default part of
name server boot files along the lines of RFC1912, Section 4.1.

Stepping back for a moment, I think we are failing to note
that this is *not* just an IPv6 problem.

The increasing volume of root queries for RFC1918 PTR maps is
presumably related to the increasing number of hosts on the
internet behind NAT boxes.  We should expect internet growth
to continue and therefore that the number of bogus IPv4 queries
will also continue to grow unless some action is taken.

We should get the IPv6 case right.  However if we believe that
dual stack is the likely IPv6 deployment scenario we are only
addressing well less than half the immediate problem.

- aidan

#----------------------------------------------------------------------
# To unsubscribe, send a message to <dnsop-request@cafax.se>.