To:
ngtrans@sunroof.eng.sun.com, namedroppers@ops.ietf.org, ipng@sunroof.eng.sun.com, dnsop@cafax.se
From:
"D. J. Bernstein" <djb@cr.yp.to>
Date:
14 Aug 2001 02:38:06 -0000
Automatic-Legal-Notices:
Copyright 2001, D. J. Bernstein. My transmission of this message to you does not constitute a copyright waiver or any other limitation of my rights, even if you have told me otherwise.
Content-Disposition:
inline
Sender:
owner-dnsop@cafax.se
Subject:
Re: (ngtrans) Joint DNSEXT & NGTRANS summary
I have a DNS server, a DNS cache, and a DNS client library. Current
situation:
(1) The server talks to caches only through IPv4. It can provide AAAA
and ip6.int records, as per RFC 1123 section 6.1.3.5, but the
input format is not as streamlined as the input format for A and
in-addr.arpa.
(2) The cache talks to servers and clients only through IPv4. It can
forward AAAA records, as per RFC 1123 section 6.1.3.5, but it
does not use them to locate servers.
(3) The client library talks to caches only through IPv4. It doesn't
include any functions related to IPv6 addresses.
Plans:
(1) Server: Talk to caches through IPv6. Include streamlined syntax
to generate AAAA/ip6.int.
(2) Cache: Talk to servers and clients through IPv6. Use AAAA records
to locate servers.
(3) Client library: Talk to caches through IPv6. Incorporate AAAA
lookups, falling back to A. Incorporate ip6.int lookups.
Things that I do _not_ plan to do:
(1) A6, DNAME, bit labels, and ip6.arpa in the server.
(2) A6, DNAME, bit labels, and ip6.arpa in the cache.
(3) A6, DNAME, bit labels, and ip6.arpa in the client library.
I want it to be crystal clear to the users that they have to set up AAAA
records and ip6.int. They can't expect to be reachable if they use A6 or
DNAME or ip6.arpa.
---Dan