RE: Comments: draft-savola-v6ops-transarch-01.txt

["Michael Mackay" <m.mackay@lancaster.ac.uk>]

Hi,

>> might it be useful to discuss these assumptions before moving on to
>> the architecture...
>
>I think this has improved; a new section has been added before 2.1
already 
>in -02 version.

Yup, I agree with 2.1 but it would be good to try to expand on it and
formalise this a bit more, if not here then somewhere. I see this as
potentially being a quite important bit of the draft and while you
certainly can't cover the whole thing in its entirety or present
complete arguments either way, there should be something that makes the
reader aware of a set of the main assumptions of high-level
transitioning deployment. 


>> How about introducing the idea of service provisioning here, not sure

>> if it would count as a general principle
>>   Services to be deployed
>>   Behaviour/performance expected
>
>Hmm.. could you elaborate a bit more?  did you have specific ideas in
mind 
>how to integrate it to the rest of the document?  That is, what level
of 
>services you're referring to?  The expected performance etc. in
realistic 
>terms might also be difficult to estimate..

Absolutely, it certainly isn't realistic to go into any sort of detail
about the actual performance you can expect but surely the services you
want to deploy over the transitioning architecture would have a
relatively large affect on the architecture you deploy. Both in terms of
'performance' characteristics (bandwidth, scalability, etc) and services
(protocols, etc) you want it to support.
I lumped them together (though they probably should be separate) to
indicate that what you operate over/expect from the architecture would
have a bearing on what you deploy in the architecture itself (and so
should count as a 'principle'?).  
    

>> Might be good to include some more discussion of the 'starting point'

>> of the transitioning, e.g.  IPv4 w/wo NAT, dual stack (various
flavours) 
>> or new (IPv6 only?)
>
>Could you elaborate a bit (please check -02 version btw, this has
improved 
>as well!)?  There *maybe* could be some brief description without
starting 
>scenarios, like v4 w/ NAT or v4 w/o NAT, but otherwise I think it has
been 
>covered..

Ah, sorry about the mix-up. I had the -01 version (that expires April
04) and hadn't seen that -02 was out. That's fine.


>> Definitely, I think the idea of discussing responsibility for the
>> provisioning of transitioning is important...
>
>Yep.. I'm not sure how to proceed from there, though.  If you have 
>thoughts how to make these issues more concrete (not just questions
here 
>and there, and discussed a bit here and there) I'm all ears :-)

Not sure how to approach this as it's something I think that needs to be
discussed and perhaps some consensus reached on the list.  Other than
emphasising that the impetus right now lies with early adopters (which
you've done) and suggesting how this might pan out in the future... 
dunno :-)
In an ideal world I suppose everyone should be responsible for their own
little bit of the transitioning scope within their boundaries, from
local administrators on their own networks to service providers and
beyond. Defining who should do what and what they should then provide to
their users might be one way to start thinking about it. What do you
think?   


>> I think this is a good draft and something that should be discussed 
>> here sooner rather than later, however it's perhaps a bit too general

>> at this stage and doesn't actually specify an architecture as much as

>> put forward some useful guidelines.
>
>I'm trying to avoid preaching "gospel", giving folks thoughts is more 
>important :-)
 
fair enough.


>> Another point is that while now it is probably best to deploy dual 
>> stack with limited IPv6, it might be an idea to outline how this is 
>> likely to change in the future.
>
>True..

I suppose it really depends upon what you see as the scope for the
draft. I agree with the phases outlined ranging from IPv4 only through
various flavours of dual stack to IPv6 only and the issues raised in
each, but it would be nice to give a clearer idea of the
preconditions/motivators for each phase and as such the prompts for the
next shift along i.e. from dual stack v4/6 to IPv6 only. This might help
readers to understand where we are now and where we are going/when we
can expect to get there.



Sorry about the mix-up, hope this clarifies things a bit.
All the best, 
Michael