[ipv6hackers] Dynamic prefixes & privacy (was: IPv6 prefix changing)

[Fernando Gont]

On 03/12/2012 04:15 AM, Alex List wrote:
> I know that this discussion is not new [3], but to me the problem in
> Germany seems to be more related to the fact that, without dynamic
> prefixes, a residential customer would be obliged to use a sort of ID
> he/she cannot change. Tracking is possible anyway. The problem is to
> what extent one can influence that. In Germany there is (imho rather
> utopic) the "law"/concept of informational self-determination [1]. A
> fixed ID in the Internet can bring back the memories of [2], a very
> sensitive topic, specially for older generations.

I think one should draw a clear line between an address, and an identifier.

Addresses will always disclose your *position* in the network. They are
location-dependent identifiers (hence call "addresses").

Requiring providers to change the subnet bits for this purpose (which I
assume it is what is being discussed here, since you're talking about
"dynamic prefixes") is a bit nonsensical, since either:
a) Changing those bits does not help much in terms of privacy, or,
b) it requires your provider to actually (topologically) move your node, or
c) All of the above :-)


An entirely different question is that about how the Interface-ID is
selected. In that request, temporary addressses (RFC 4941) make sense,
and also the current MAC-derived addresses should also be replaced by
the scheme proposed in
<http://tools.ietf.org/id/draft-gont-6man-stable-privacy-addresses-00.txt>.

Thanks,
-- 
Fernando Gont
SI6 Networks
e-mail: fgont at si6networks.com
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492