Re: REVIEW NEEDED: draft-ietf-v6ops-ent-analysis-00.txt (fwd)

[Tim Chown <tjc@ecs.soton.ac.uk>]

On Thu, Oct 14, 2004 at 09:06:43PM +0300, Pekka Savola wrote:
> 
> I see a problem with not discussion the omitted ones at all.  You've
> been through the though exercise of going through all the
> combinations, and using some methodology, picked the ones that you
> felt were interesting.  I think that methodology needs to be carefully
> described because otherwise the people who might have different 
> assumptions than you (on what they consider important or not) probably 
> arrive in a different result.

The reality of course is that ent-scenarios was written before we had
the very thorough (and useful!) matrix presented to the WG.   Thus we
have ent-scenarios with three broad scenarios (two of which are
dual stack and IPv6 dominant), and now the added "complexity" of what
some would call "corner cases".   The thing is, more than one of these
cases are being deployed now...
 
> My main concern was on v6 ISPs, though I'd be a bit concerned of the
> v6-only -> v4-only app scenario.  That would seem to be like generic
> v6-only + NAT-PT direction, which folks may or may not agree with.  
> But that can certainly be discussed.

But this is a real scenario that needs to be addressed.   You either have
translation (at layer 3/4/7)  between v4 and v6 only networks, or on
the v6 network you run v4+NAT in parallel for those users to get *out*
of their network to run legacy web/ftp/mail apps.   I'm sure we'll see both
cases in the wild, in not insignificant volume.
 
> > Not sure I agree.  V4 or v6 transport is a choice we assume dual stacks?
> 
> I meant, v6 transport for DNS is not required when you run dual-stack.  
> It's nice, of course, but not required.  It's only required with
> v6-only.

Or if you want a dual-stack network (infrastructure including routers,
links and services) into which a v6-only node can connect and run with all
services available.   In our deployment, that's our goal.
 
> > > ==> it would be worth mentioning (at least as a problem, if not 
> > > further
> > > discussed) in a new paragraph how one would configure such proxies or 
> > > translators on the v6-only nodes.  Manually?
> > > Using yet-to-be-defined DHCPv6 options?  Using unspecified means?
> > 
> > Hmmmm to some degree yes but we don't want to own that one in this spec
> > we may need some WG help on this one.
> 
> Well, at least it could be noted as an open issue, even if nothing 
> concrete is done or decided in this document. (In any case, this doc 
> should not give "normative" guidance on which way to go, because we'll 
> likely need IETF process to run on this).

Fair point.  It's an important part of the translation vs dual-stack puzzle
referred to above.
 
-- 
Tim