ipv6/v4 naming nomenclature [Was: Apple Air...]

Jeroen Massar jeroen at unfix.org
Wed Sep 19 23:13:27 UTC 2007

Barrett Lyon wrote:
> On Sep 18, 2007, at 1:30 PM, David Conrad wrote:
>> On Sep 18, 2007, at 5:45 AM, Jeroen Massar wrote:
>>> Please please please, for the sake of a semi-'standard', please only use
>> What RFC (or other standards publication) is this documented in?
> Where did the www.ipv6 and www.ipv4 "standard" come from?

Note my clear use of >>semi<< and >>'standard'<< and mind the quotes.
Also note that for instance an IETF "standard" is only a standard when
something in very common use for quite some time. Maybe this would be a
good one to jot down as an Informational/BCP kind of document.

It is something which is in use by a lot of sites who have been enabled
with IPv6 for about the last 10 years and needed a way to distinct IPv4
and IPv6 variants of their hosts. Remember, before that people on NANOG
started noticing the existence of IPv6 in the last few months, and
before we had the 6bone with 3ffe::/16 there was also a 6bone with
5f00::/8 (RFC1897, which I really had to look up as my bear is not that
long ;), oddly enough that is even before most people even had internet
or knew that it existed.

> As for
> end-users such as normal non-network people, having a standard that adds
> more characters than necessary (that eventually may become arbitrary)
> seems rather silly.

It is not meant to be used by end-users. Those should simply
Google/Yahoo/Baidu for the description of the site and get the content
and not be bother with remembering hostnames, let them use bookmarks or
something, then they at least won't be caught by typo-squatters who are
dominating the DNS system.

It is only a semi-'standard' which is in use by network operators who
didn't want to remember the AAAA or A for a hostname while being able to
choose a protocol to ping/telnet/ssh/etc as there was a time when we
already had IPv6 but some tools (yes, including PuTTY ;) did not allow
selection of either IPv4 or IPv6 protocols. But also allowing the hosts
to have an AAAA + A so that a tool could pick the protocol that was
available. Sometimes you want to select that thus using:
 hostname.example.com      A + AAAA
 hostname.ipv6.example.com AAAAA
 hostname.ipv4.example.com A
solved that problem without having to add support for a '-6' or '-4'
switch for IPv6 and IPv4 respectively into all tools that one uses.
Some code doesn't come with source and some people don't want to patch
it up so this is a perfect way to solve it.

As mentioned above, this is for people who want to pick the version of
IP protocol used. End-users don't even know what "IP" is, and they also
should not know and they definitely should not have to care. As such,
when you think that your site is 'working fine' over IPv4, then add an
"A" record to it, when you think that IPv6 is fine, add an AAAA record,
otherwise, have a trick like http://www.braintrust.co.nz/ipv6wwwtest/
and test if your users can reach the IPv6 variant or not. Most likely
they won't have IPv6 enabled yet, if they have then great :)

> Why wouldn't w4.<domain> or w6.<domain> suffice for
> this purpose rather than making it overly scientific?

It would suffice, it just is not what is in use. Also some people like
to actually name OTHER things than their 'www' with it. The trend for
that seems to be more that you can ignore the 'www' portion anyway and
just http://youtube.com or http://google.com work.

Also, I know a certain company using 'w3' for intranet-only websites,
while using 'www' for internet websites.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 311 bytes
Desc: OpenPGP digital signature
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20070920/51a87110/attachment.sig>

More information about the NANOG mailing list