A DNS question re 6to6/IPv6 host IN A records.

[Pekka Savola]

btw, I suggest you add an emptyy line between paragraphs, it makes reading 
easier.

On Fri, 19 Apr 2002, Jeroen Massar wrote:
> Pekka Savola wrote:
> 
> > On Fri, 19 Apr 2002 itojun@iijlab.net wrote:
> > > >In the forward/reverse zones on a 6to4 setup, should I have
> > > >nanguo          IN A                203.1.96.5
> > > >nanguo-v6      IN AAAA         2002:cb01:6005:2::1
> > > >or
> > > >nanguo          IN A                203.1.96.5
> > > >nanguo          IN AAAA         2002:cb01:6005:2::1
> > > >When referring to the particular host ?
> > > >Either works - but which is ... errr... correct?
> > > 
> > > 	i recommend the latter, definitely.  with the latter 
> > you will be able
> > > 	to transition to IPv6 much smoother.
> > 
> > That is true, but it may have it's drawbacks.  Often, still, IPv6 
> > connectivity is worse than with IPv4.  People who are 
> > dual-stack will use IPv6 when trying to reach 'nanguo'.  It may be
> > more unoptimal yet.

> Sorry to say it but I really think it's a load of B.S.... in my opinion
> anyways.

You're entitled to be wrong ;-)

> Most hosts I 'use' most of they day and that are IPv6 connected are as
> close as when I would use IPv4.

As do I.

> I use IPv6 transparently fortunatly so I usually don't even notice the
> difference between IPv6 and IPv4.

As do I.

> Remote hosts (non-european :) though are flaky sometimes. Certainly this
> would improve very much
> when all those tunnels crossing multiple AS's dissappear, it will take
> some time but it will come one day ;)

Sure, but we're talking about what makes sense *today*.  Not in 2 years 
:-).

> KAME is about 300ms 'away' from Holland most of the times in both IPv4
> and IPv6, so I wonder why IPv6 has 'drawbacks' over IPv4.

KAME is but one site.  There are others.  Much less technically capable 
too (e.g. ones getting tunnels from topographically far away, making the 
latency worse).

> > For conservative IPv6 adoption, I recommend the former (at least
> first).
> > For more radical IPv6 adoption, and for non-production services, the 
> > latter is usually more suitable.
>
> The second is certainly production capable. 

I don't agree.  There may be differences what one means with 'production' 
though.  Personally, if I had a power to switch on IPv6 on www.google.com 
hosts, I would only do it by adding www.ipv6.google.com: NOT with 
www.google.com.  People who are afraid of degrading service and it costing 
real money are reluctant.

An example of potentially technically capable stuff: IPv6 service of 
playground.sun.com/ipv6/ was down/out-of-sync some time ago.. and that 
isn't even a "production" site.

> Why should it be
> "non-production" anyways.

Any client node implementing IPv6 gets potentially worse service.

> Slide 50 shows a pragmatic projection of IPv6 deployment with the US
> tagging behind Asia for about 2.5 years and 1.5 years behind Europe!
> I sure hope that changes quite soon over there.

Similar slides have been seen before, e.g. at IPv6 deployment conference 
(ask Pim.)

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords