[99314] in North American Network Operators' Group
Re: Apple Airport Extreme IPv6 problems?
daemon@ATHENA.MIT.EDU (John Curran)
Mon Sep 17 17:29:10 2007
In-Reply-To: <2d106eb50709171347p893a8bke0ea4c00916657f9@mail.gmail.com>
Date: Mon, 17 Sep 2007 17:18:07 -0400
To: "Martin Hannigan" <hannigan@gmail.com>,
"Iljitsch van Beijnum" <iljitsch@muada.com>
From: John Curran <jcurran@mail.com>
Cc: "Barrett Lyon" <blyon@blyon.com>, nanog@merit.edu
Errors-To: owner-nanog@merit.edu
At 4:47 PM -0400 9/17/07, Martin Hannigan wrote:
>On 9/17/07, Iljitsch van Beijnum <iljitsch@muada.com> wrote:
>> On 17-sep-2007, at 19:06, Martin Hannigan wrote:
>>
>> > Getting back to my original discussion with Barrett, what should we do
>> > about naming? I initially though that segregating v6 in a subdomain
>> > was a good idea, but if this is truly a migration, v4 should be the
>> > interface segregated.
>>
>> For debugging purposes, it's always good to have
>> blah.ipvX.example.com, but the real question is: do you feel
>> comfortable adding AAAA records to your production domain names?
>> Although I've been running that way for years and I've had only one
>> or two complaints during that time, I can see how someone could be
>> worried about reduced performance over IPv6 (it's still slower than
>> IPv4 a lot of the time because of tunnel detours etc) or even
>> timeouts when advertised IPv6 connectivity doesn't work for someone,
>> such as a Vista user with a public IPv4 address behind a firewall
>> that blocks protocol 41.
>>
>> Then again, I'm guessing that few people type www.ipv6.google.com
>> rather than www.google.com. And with stuff like mail, where you set
>> up the server names once and forget about it, it's even worse.
>>
>
>
>I see. There isn't really an answer. :-) That's what I am getting at.
>Not to suggest that this is your responsibility, it's not - it's ours.
>
>For now, I'm going to try the unique A/AAAA and segregate the answers
>by protocol and sub domain the v4 traffic since it's a migration "to"
>v6.
>
>
>-M<
