Re: [DNA] CPL & Solution Framework Issues

[Greg Daley <greg.daley@eng.monash.edu.au>]

Hi JinHyeock,

JinHyeock Choi wrote:
> Dear DNA WG 
> 
> At San Diego meeting, Erik Nordmark and I will present the following drafts. 
> 
> 1. DNA with unmodified routers: Prefix list based approach
> 2. DNA solution framework
> 
> To make the best out the discussions, we present the WG, 
> for CPL and solution frame, 
> 
> 1) what issues we'd like to fix and
> 2) what decisions we'd like to make 
> 
> at the f2f discussions. 
> 
> 1. DNA with unmodified routers: Prefix list based approach
> 
> We wish DNA WG to accept the draft as a WG item for DNA BCP work 
> (after the internal review if necessary). 

I think that the work is necessary for a BCP but is it sufficient
to describe DNA?

We can talk about that referring between Sathya's and the CPL
draft in the session (and before).

Possibly the CPL draft could stand on its own and have some
additional information about validity of NC entries, authorization
dad requirements for DNA nodes etc provided in another document
which references CPL.

Otherwise, if we're only allowed to put out one document, CPL should
compose at least a part of that.

At the moment I think the draft is fairly well written, but the
on-list discussions may make some change to the composition.
The idea is fairly firm though.

> 2. DNA solution framework
>  
> Most of all, we wish that WG will reach consensus on 
> Sec 2 Basic Assumptions and Sec 3 DNA Solution Sketch. 
> 
>    2.  Basic Assumptions  
>      2.1   DNA solution based on link identity detection   
>      2.2   Checking for link change with Link Identifier  
>      2.3   RA message optimized for DNA 
>      2.4   Quick delivery of an RA  
> 
>    3.  DNA Solution Sketch  
>      3.1   Solution components  
>      3.2   Solution procedure 
>      3.3   Work items 
> 
> It's our wish that DNA WG will agree on the 'Basic Assumptions' 
> and 'DNA solution Sketch' and decide to work in that direction.  

I think that the first is almost essential.
The second may be possible, but relies upon the other.

> But the framework document (and 5 minute presentation) might 
> not be sufficient to convince WG of the necessity of LinkID.
> 
> At least, we'd like DNA WG will clearly understand and explicitly accept 
> 
>      2.1 DNA solution based on link identity detection 
> 
> Your feedback will be appreciated. 

We may take a little longer than just the meeting for that,
but I'd hope that we get closer to consensus on that.
If we keep discussing things on the list, we may come to a
similar conclusion.

Of course all decisions in the meeting need to be confirmed on
the list.

As we've discussed with CPL, there are some quite tricky
requirements (Link Hints, or other information) which may be
needed to provide identification of the link.

If we have an explicit LinkID, some of the guesswork disappears.

If CPL can be found to provide sufficiently good link identification
itself, we may not see an explicit link identifier in the solution.

Greg