[arin-ppml] The non-deployment of IPv6

[Davis, Terry L]

It is interesting that we are now finally worried about the root causes of the v6 "non-deployment".  Some of us have been worrying this for years...

To me it comes down to a few points:
-          Cost of conversion!  We need a truly seamless v4 to v6 interface to reduce implementation costs and risks.  As it is, every year that goes by without v6 implementation increases its implementation cost and conversion risks; thus adding more resistance every year to conversion.

-          We need to get IPv6 addresses in the hands of the developers; especially the small greenfields.  Perhaps our past address allocation policies just might have something to do with the problem.  In the mid 80's to early 90's, an email request with almost zero justification would get 50 or 100 class Cs or a class B without regards to the size of the entity or what industry it was part of.  Are we still too restrictive for the greenfields and new technology developers?

-          The vendors all need to show, advertise, and evangelize that their v6 products coexist harmlessly with existing v4 infrastructures so that a network owning organization (agency, enterprise, etc) will see no risk in adding v6 support to their networks.

-          New products and new product versions need to support both v4 and v6 seamlessly. A few folks have the right idea but not enough.

Take care
Terry

________________________________
From: arin-ppml-bounces at arin.net [mailto:arin-ppml-bounces at arin.net] On Behalf Of Durand, Alain
Sent: Tuesday, December 08, 2009 11:29 AM
To: matthew at matthew.at; Ted Mittelstaedt
Cc: arin ppml
Subject: Re: [arin-ppml] The non-deployment of IPv6

On 12/8/09 2:01 PM, "Matthew Kaufman" <matthew at matthew.at> wrote:
Ted Mittelstaedt wrote:
>
> I agree that content providers NEED to be dual-stacked first,...
And where will they be getting the IPv4 for that side of "dual-stacked"
from after runout?

===> This is where a technology like dual-stack lite comes to play. See
http://www.ietf.org/id/draft-ietf-softwire-dual-stack-lite-02.txt

And if that's via a mechanism that actually works well enough, then why
would we need IPv6?

===> Because a packet that flows natively over IPv6 is a packet that does not have to go through the IPv4 NAT infrastructure.
In other words, the more v6 traffic, the less we spend on NATs.

     - Alain.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20091208/b2b4e24d/attachment.htm>