Re: draft-bourbaki-6man-classless-ipv6-00

[Brian E Carpenter <[e-mail]>]

On 13/06/2017 14:30, Manfredi, Albert E wrote:
> I should add,
> 
> The IID length used by all SLAAC hosts on a link would be the same, because presumably, they all receive the same RAs. With the same RAs, their IIDs would have to come out to be the same length.
> 
> And, most of this change doesn't even apply to RFC 4862. 

It applies to the fact that the IID length must be known before forming
a LL address, and that it must be the same length as implied by a
*later* RA/PIO with A=1. So actually I think there is quite some
work to be done on RFC4862, and it isn't clear to me that it could
be backwards-compatible.

Also see Gengyu Wei's message: when devices roam, what happens if
the LL IID length for a given interface is supposed to change?
DIP switches again?

I don't think we're going to see a different IID length for any
link type that's already using 64 today; too many things would
have to change.

> For the most part, it applies to RFC 2464-bis. The main points of RFC 4862 remain the same.

It doesn't matter. If it isn't backwards compatible with deployed
code, IMHO it isn't going to happen on Ethernet or Wi-Fi.

   Brian