Re: [DNA] Updates to dna-protocol3

[Greg Daley <gregd@research.panasonic.com>]

Hi Brett,

Thanks for the changes.

Brett Pentland wrote:
> Before submitting the draft to become draft-ietf-dna-protocol-00.txt I
> have some updates to reflect the feedback that we've had so far.  The
> update can be viewed at:
> 
> http://www.ctie.monash.edu.au/dna-dt/draft-ietf-dna-protocol-pre00.txt
> 
> And a diff from protocol3 can be seen at:
> 
> http://www.ctie.monash.edu.au/dna-dt/draft-ietf-dna-protocol-pre00-from-draft-pentland-dna-protocol3-00.diff.html
> 
> 
> Does anyone have any objections to the amendments made so far?  Do they
> break anything?  Positive ACKs are ok too. :)
> 
> There's a few issues that have come up while making the edits.  They
> don't necessarily have to be resolved before submission of the -00, but
> some input would be useful.
> 
> - Should the explicit linkid always be in the LPO or can it be a flag in
>   a PIO (the latter is in the draft)?

It is a significant (16 octet) cost saving to make use of a flag in the
PIO to explicitly mark a link identifier.  The processing is more
complicated in that PIOs and Landmark options have to be checked until
the identifier is found.

This per prefix processing is required if there's no identifier (using
CPL) though, so that's not a major problem.


> - What order do the hosts treat the indications (land mark option,
>   prefixes, linkid in the draft at present)

I'd be interested in finding out what others think.
My personal thought was that the completeRA flag trumps all.

> - Given that we now have an explicit linkid, should we get rid of all of
>   the flags in the landmark option - the no flag is not really needed
>   and since taking that out would leave only one, we can probably get
>   rid of that as well - LO in RA = yes (at present the flags are still
>   there)

it doesn't make the option any smaller to remove. perhaps we can leave
it for future extension?

Greg