Re: draft-gont-6man-managing-privacy-extensions-00.txt

[Fernando Gont <[e-mail]>]

On 16/03/2011 03:30 a.m., Christian Huitema wrote:
>> Then what's all this controversy with
>> draft-gont-6man-managing-privacy-extensions? :-) -- That aside,
>> there have been quite a few publications asessing the real
>> "privacy" provided with the so-called privacy-extensions....
> 
> Using randomized host identifiers is way more private than sticking a
> MAC address in the IPv6 address. 

Agreed. Also in some scenarios privacy extensions may give a false sense
of privacy (see e.g. some of the pointers in our I-D).


> In fact, rather than your draft
> proposing to not use privacy addresses, we should pursue the
> deprecation of using EUI-64 in addresses.

Our draft is not meant to propose "not to use privacy addresses" -- as
noted a few times, already, the proposed mechanism could be used to turn
"privacy addresses" on for some systems that have decided not to enable
them by default (e.g., FreeBSD).

I realize that the case is made only for those scenarios in which an
admin might want to disable privacy addresses, though.


> The worst part of your draft is that , if we published it, it would
> give the impression that for the IETF, EUI-64 addresses are more
> secure than privacy addresses. This is obviously false.

We didn't argue that "privacy addresses" should be off by default.
Actually, I'd probably argue the opposite: for the general case they
should probably be "on"... but the admins should have the means to
manage how address are selected (whether privacy addresses, EUI-64, or
anything else).

Thanks,
-- 
Fernando Gont
e-mail: [e-mail] || [e-mail]
PGP Fingerprint: 7809 84F5 322E 45C7 F1C9 3945 96EE A9EF D076 FFF1