>________________________________
> From: Christian Huitema <huit...@microsoft.com>
>To: Brian E Carpenter <brian.e.carpen...@gmail.com>; Andrew McGregor
><andrewm...@gmail.com>
>Cc: "ipv6@ietf.org" <ipv6@ietf.org>
>Sent: Saturday, 6 April 2013 2:35 AM
>Subject: RE: link local address handling while changing hw address of interface
>
>> At layer 2, yes, but you haven't done anything at layer 3, and NUD/ND will
>> pick up the new layer3:layer2 relationship.
>
>There is also a privacy angle to this. Suppose that a visitor to a café
>changes their MAC address in order to avoid tracking. Keeping the same IP
>addresses would somewhat defeat the purpose. Of course, there is always a way
>to use "ipconfig renew" or some equivalent command, but you have a race
>condition between "starting to use the new MAC" and "removing the old
>address." The only way to address that condition is to make MAC change and
>address renewal an atomic operation.
>
>
>Some of the modern wired network cards can support multiple unicast
>destination addresses, mainly to support virtualisation if I understand
>correctly, although I'm not sure about any 802.11 cards. An interesting idea
>would be to utilise this capability to cover a transition period between an
>old MAC address to a new one for privacy reasons. Alternatively most, if not
>all NICs could be put into promiscuous mode during this transition period to
>support receiving traffic for the old as well as new MAC address.
>
>Given the recent use of Wifi MAC addresses to supposedly uniquely track
>visitors to a store, and the far worse than expected level of MAC address
>global uniqueness that HD Moore discovered, I'd be inclined to start looking
>at periodically changing my MAC address in public arenas.
>
>
>Regards,
>Mark.
>
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------