Re: Multihoming Issues

"Jim Fleming" <JimFleming@ameritech.net> Wed, 04 September 2002 00:02 UTC

Received: from loki.ietf.org (loki [10.27.2.29]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id UAA17725; Tue, 3 Sep 2002 20:02:25 -0400 (EDT)
Received: (from adm@localhost) by loki.ietf.org (8.9.1b+Sun/8.9.1) id TAA05716 for ietf-outbound.09@loki.ietf.org; Tue, 3 Sep 2002 19:40:00 -0400 (EDT)
Received: from ietf.org (odin.ietf.org [10.27.2.28]) by loki.ietf.org (8.9.1b+Sun/8.9.1) with ESMTP id TAA05685 for <ietf-mainout@loki.ietf.org>; Tue, 3 Sep 2002 19:37:28 -0400 (EDT)
Received: by ietf.org (8.9.1a/8.9.1a) id TAA17230 for ietf-mainout@loki.ietf.org; Tue, 3 Sep 2002 19:35:53 -0400 (EDT)
X-Authentication-Warning: ietf.org: majordom set sender to owner-ietf@ietf.org using -f
Received: from mailhost.chi1.ameritech.net (mailhost1-chcgil.chcgil.ameritech.net [206.141.192.67]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA17220 for <ietf@ietf.org>; Tue, 3 Sep 2002 19:35:48 -0400 (EDT)
Received: from repligate ([67.37.176.103]) by mailhost.chi1.ameritech.net (InterMail vM.4.01.02.17 201-229-119) with SMTP id <20020903233722.YUXA18992.mailhost.chi1.ameritech.net@repligate>; Tue, 3 Sep 2002 18:37:22 -0500
Message-ID: <043001c253a2$de8041f0$8c56fea9@repligate>
From: Jim Fleming <JimFleming@ameritech.net>
To: Michel Py <michel@arneill-py.sacramento.ca.us>, Caitlin Bestler <caitlinb@RP.ASOMI.NET>
Cc: 'The IETF' <ietf@ietf.org>
References: <2B81403386729140A3A899A8B39B046405E2CF@server2000>
Subject: Re: Multihoming Issues
Date: Tue, 03 Sep 2002 18:37:29 -0500
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Content-Transfer-Encoding: 7bit
Sender: owner-ietf@ietf.org
Precedence: bulk
Content-Transfer-Encoding: 7bit
X-Loop: ietf@ietf.org
Content-Transfer-Encoding: 7bit

----- Original Message ----- 
From: "Michel Py" <michel@arneill-py.sacramento.ca.us>
> 
> > This thread started with exactly such a question
> > being raised, but the rationale on how DNS *could*
> > be optimized for IPV6 was not spelled out.
> 
> If you can't relate to a specific protocol, there is little point
> spelling out how DNS could be optimized in the abstract.
> 

It might be easier if you start with an Architecture.
128-bit DNS is useful in extending the ability to set more
bits in the 160-bit IPv4 header (plus the UDP or TCP headers).
The current 2002 "version" of 128-bit DNS is rather simple.
People start on the left with 2002:[IPv4]:* where IPv4 is one of
their site-ids. That allows an IPv4 header to be constructed with
that as the Destination Site-ID. For multi-homing, a node has to
be able to determine a Source Site-ID so that packets can find
their way back. Historically, the trivial 32-bit DNS has simply
been a name to A record mapping service, with only the destination
address in the 32-bits of the A record. The source knows its
address, or is supposed to.

One could imagine designing next year's version (2003) to be
2003:[IPv4src]:[IPv4dst]:*. That would allow the DNS to store binary
information on site-to-site prefixes as a name. It is not clear people
would want to have a lot of names to describe the *relationships* between
nodes, as opposed to just the location of the nodes. Also, that would take
80 bits (16+32+32) of the 128-bit DNS AAAA record, leaving 48 bits
for a persistent address. That makes it hard to also include a 16-bit Port
Number (for UDP | TCP) and have more than 32-bits left for the IPv4++
extended address bits. If one only cared about the legacy IPv4 transport
then they could use the 48-bits for one more persistent IPv4 address and
16-bits for the Port Number.

This would be....2003:[IPv4src]:[IPv4dst]:[IPv4]:[Port]...filling the 128-bits.
IPv4src of 0 in the 128-bit DNS could mean "any" or the default site-id.
This 2003 version could be designed for people who want to take an
Intranet and expose it to the global transport. Adding modules to do this with
the new stack on Linux is rather easy. Many more of the 160-bits in the classic
IPv4 header are now coming into play for routing, and the 128-bit DNS can
be used to set those bits.
http://www.netfilter.org

In 2004, someone may have a different version to suggest...

Jim Fleming
2002:[IPv4]:000X:03DB:...IPv8 is closer than you think...
http://ipv8.dyndns.tv
http://ipv8.yi.org
http://ipv8.dyns.cx
http://ipv8.no-ip.com
http://ipv8.no-ip.org
http://ipv8.no-ip.biz
http://ipv8.no-ip.info
http://www.iana.org/assignments/ipv4-address-space
http://www.ntia.doc.gov/ntiahome/domainname/130dftmail/unir.txt