Re: Software to test MAC address privacy

[Tom Gundersen <teg jklm no>]

Hi Robert,

On Fri, Jul 18, 2014 at 3:06 AM, Robert Moskowitz <rgm htt-consult com> wrote:
> Greetings from IEEE 802 plenay in San Diego.  We are winding down, but
> Monday night we had a talk on Pervasive Surveillance:
>
> https://mentor.ieee.org/802-ec/dcn/14/ec-14-0043-00-00EC-internet-privacy-tutorial.pdf
>
> I discussed this with the 802 chair and presentation moderator, and we are
> looking to see if we can actually test the consequences of using random
> local MAC addresses.  The idea is to have an opt-in SSID at future 802
> meetings, and perhaps at the IETF as well (same network support company)
> where only random local MAC addresses are allowed and then to see what
> problems occur (DHCP, ARP tables, bridging tables, etc.).
>
> So we (those of us that want to figure this out to see if it is worth doing)
> are looking to the OS providers to help.  I have been tasked with reaching
> to the Linux community as I run Fedora.
>
> The thought is the MAC address is temporarily overwritten with a local MAC
> random address.  This address should be changed with some periodicity.

Recent versions of udev has the possibility of optionally set a random
mac address on every boot (or device hotplug) [0]. However, it will
not change the mac address at runtime (which appears to be what you
want).

> We have not worked out this part yet.

As I have advocated in the past, I think it may make sense to set a
random MAC address per SSID, so that you won't change the MAC address
whilst connected, but at the same time you can not be tracked across
SSIDs (there are still some issues to solve with that though). One
should obviously also do as OSX does and use random MAC addresses when
scanning.

Sounds like a nice initiative. Best of luck!

Cheers,

Tom

[0]: <http://www.freedesktop.org/software/systemd/man/systemd.link.html#MACAddressPolicy=>