Robert,

On Oct 8, 2010, at 2:35 PM, Robert Cragie wrote:

> Just to reiterate on ND: The creation of an NCE to hold the unconfirmed IP/LL 
> address of the joining node does not make sense for the reasons Colin has 
> clearly outlined. It also contradicts generally the usage of options to carry 
> information where statelessness is required, e.g. RS/RA. Therefore, as I have 
> said many times before, it makes more sense to carry this (these) addresses 
> in the options.


Please go back and check http://tools.ietf.org/html/draft-ietf-6lowpan-nd-10. 
Notice that in this version there was no tentative NCE made while multihop-DAD 
was being performed. Instead all the needed information was carried in an ARO + 
SLLAO combination, and the NCE was made only after a successful response came 
back from the 6LBR. Then the WG asked us to change that, and instead make a 
tentative NCE (which appeared in -11).

My person view is that if the WG does not see a need for the tentative NCE, and 
just finds it confusing, then I would have no problem removing that and instead 
carrying the L2 address with the ARO similar to what was done in -10. Notice 
that this doesn't enable the 6LR to detect a duplicate while an address is 
being checked by the 6LBR, but it will be detected by the 6LBR instead. A rare 
case anyways, so I don't see it as an issue.

Note that this issue is orthogonal to the NS/NA+ARO exchange between the 6LN 
and 6LR. 

Zach

-- 
Zach Shelby, Chief Nerd, Sensinode Ltd.
http://zachshelby.org  - My blog "On the Internet of Things"
http://6lowpan.net - My book "6LoWPAN: The Wireless Embedded Internet"
Mobile: +358 40 7796297

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan

Reply via email to