Re: [OPSAWG] WGLC on shared transition space

[Matthew Kaufman <[e-mail]>]

On 8/5/2011 10:35 AM, Chris Grundemann wrote:
>> Right. And the use cases should talk *exclusively* about enabling NAT444 so
>> that dual-stack w/IPv6 + NAT444 is possible, *not* about several other ways
>> this shared block can be (mis)used to extend the lifetime of
>> IPv4 services.
> <...>
>> Then modify this document to restrict the use to exclusively NAT444.
>> Otherwise the resulting RFC will create a mixed message, and those
>> corporate IT departments will be able to point right at those sections when
>> arguing with you about what your NAT did to them.
> I am not sure we are reading/discussing the same document, just to be sure here is the bit I find relevant to your statement (from the ID in WGLC - http://www.ietf.org/id/draft-weil-shared-transition-space-request-03.txt):
>
> 4.  Shared Transition Space
>
>     This document proposes the assignment of a /10 as Shared Transition
>     Space.  Shared Transition Space is IPv4 address space reserved for
>     Infrastructure provider use with the purpose of facilitating IPv6
>     transition and IPv4 coexistence deployment.  The requested block
>     SHOULD NOT be utilized for any purpose other than as "inside"
>     addresses in a carrier NAT environment (e.g. between the CGN and
>     customer CPE devices) or for other IPv4 to IPv6 transition
>     infrastructure.  Network equipment manufacturers MUST NOT use the
>     assigned block in default or example device configurations.
>
> I find that quite clear on what an acceptable use of this space is. The details of applicability of and justification are left out, for further debate, but the intent of this space is well defined in the draft, imho. It IS for use by service providers to facilitate transition to IPv6, it is NOT for use by anyone else (corporate IT departments included) for any other purpose.

Change SHOULD NOT to MUST NOT.

More clearly define "or for other IPv4 to IPv6 transition 
infrastructure" so that it is clear that it doesn't include things like 
CDN server IPv4 addresses  or anything else that prolongs the utility of 
IPv4 by numbering things in this space.

And if we're ever going to take the other draft 
(draft-bdgks-arin-shared-transition-space) or something like it to last 
call, it should go to last call *first* so that this draft can reference 
it. Otherwise we're going to reopen this same can of worms in short order.

Matthew Kaufman