Re: A common problem with SLAAC in "renumbering" scenarios

[Alexandre Petrescu <[e-mail]>]

Le 15/02/2019 à 10:37, Christian Huitema a écrit :
> 
> 
>> On Feb 14, 2019, at 11:30 PM, Jan Zorz - Go6 <[e-mail]> wrote:
>> 
>> /64 for each device is fine. What I'm questioning is if we really
>> need to make underlying to be something that was never meant to be
>> if L7 and above is broken? I think L7 and above needs to be fixed
>> first and then changing of addresses in transport layer can have
>> any effect. Until then - let's transport packets in a way that
>> doesn't break often and make user grumpy.
> 
> We are pretty far from the original thread, but let me say that
> looking at layers as if they were isolated does not work. Tracking
> uses all the cues that it can, from MAC addresses to GPU model. It
> will of course use 64 bit prefixes when available. That a beautiful
> index for the surveillance operations. It is pretty clear that we
> need to somehow defend against that.

Address the surveillance operations at their root would allow freedom to 
design session continuity IPv6 mechanisms safely assuming that the root 
of the surveillance operations is ok.

If possible :-)

(I can assure you that changing IP addresses all the time is a nightmare 
in implementations :-)

Alex

> 
> -- Christian Huitema 
> -------------------------------------------------------------------- 
> IETF IPv6 working group mailing list [e-mail] Administrative
> Requests: https://www.ietf.org/mailman/listinfo/ipv6 
> --------------------------------------------------------------------
>