Re: [v4tov6transition] Some opinions about establish a new WG

[Fred Baker <fred@cisco.com>]

I'll remind you, btw, that "draft-ietf-v6ops..." means that the working group was asked if they wanted to support it, and said "yes". I'm willing to talk about why it's in the WG, but - it's in the WG.

On Aug 27, 2010, at 3:29 PM, Fred Baker wrote:
> On Aug 27, 2010, at 3:02 PM, Randy Bush wrote:
>>> The continuation of IPv4 service during transition, what I call the "coexistence phase", is at this point very near term transition issue, and is the subject in part of
>>> 
>>> http://datatracker.ietf.org/doc/draft-ietf-v6ops-incremental-cgn
>> 
>> my apologies for being picky again.  but is this really about transition, or more about yet another complex solution for how to hack ipv6 through a severely damaged ipv4 network?  i.e. it's a we're-not-gonna-transition draft.  maybe we should call it ipv4 'intransigent' technology :)
> 
> That's not how I read it. I read it as pretty responsive to what the operators on v4tov6transition@ietf.org are asking: "while we deploy IPv6 with global addressing, we have an existing IPv4 business to maintain. We have issues in how we do that. How should we do that?"
> 
> It says, in essence, "run two networks in parallel". One is a vanilla IPv6 network. The other is, gak, an IPv4 network in which several instances of RFC 1918 address space are run in parallel to IPv4/IPv4 NATs within the SP. Examples of such exist in many networks; at the recent Google symposium, Verizon is quoted (I wasn't there) as saying they had 70 instances of such running in parallel. Since homes frequently use 192.168.0.0/16 or 10.0.0.0/8, I would expect ISPs might use 172.16.0.0/12 in their layer. It has all the issues of NAT'd address space, and being dual layer, has them even more so, so it's not a very wonderful service. But, for client/server applications whose servers are in the global address space and for p2p applications with supernodes and meet-me points in the global address space, it works. It permits the ISP to keep its business functional while deploying IPv6. Having deployed IPv6, and having their customers upgrade CPE software and/or hardware to be able to use it, they will have a much better network and better service. 
> 
> I opened a discussion this morning with my AD on this topic; I will argue that it is perhaps in a grey area of the charter (it is an issue related to deployment and transition, which is in scope but per the charter of less emphasis). We as a working group will probably wind up amending the charter to make this more clear.
> 
> I don't think, though, that v6ops - an operationally-focussed working group - can practically say "we're only going to look at the issues of networks that have finished deployment; we will not look at the issues of networks that want to deploy and have problems". I understand legitimate problems to be in scope, and this is a problem that they legitimately have.
> 
> In any event, once the WGLC is completed and the document is off to the IESG, we won't need to have this discussion regarding it.