Re: I-D Action: draft-ietf-6man-default-iids-04.txt

[Fernando Gont <[e-mail]>]

Hi, Christian,

On 07/06/2015 06:15 PM, Christian Huitema wrote:
> 
> From: ipv6 [mailto:[e-mail]] On Behalf Of Kerry Lynn 
> Sent: Monday, July 6, 2015 10:16 AM To: Fernando Gont Cc:
> [e-mail]; Dave Thaler; Ralph Droms;
> [e-mail]
>>> As the lead author of 
>>> https://tools.ietf.org/html/draft-ietf-6man-6lobac, I'm strongly
>>> in favor of retaining the option to specify IIDs based on
>>> locally assigned link-layer addresses, particularly for
>>> link-local addresses.
>> 
>> FWIW, this has never been under question. That is, the question is 
>> regarding the wording, not about the intent. Based on past wg
>> discussion and consensus, we all agree that there must be room to
>> go against this advice if design/engineering tradeoffs warrant
>> this.
> 
> Part of the problem is that "link layer address," as a category, is
> no well-defined.
> 
> I would prefer something stronger, like " Future specifications
> SHOULD NOT specify IPv6 address generation schemes that embed the
> underlying link-layer address in the IID. In some cases, where
> engineering constants dictate, identifiers MAY include a short
> dynamically assigned link-layer address. Notwithstanding that,
> specification MUST NOT embed a globally unique identifier such as an
> IEEE802 address or a E.164 phone number."

Two comments:

1) Getting into the underlying details can be tricky: short and
temporary identifiers might be "good enough" wrt host-tracking, but
certinaly not god when it comes to network reconnaissance. So I'd keep
things the discussion as is.. otherwise one would need to get into a
discussion ov every sort of identifier that might be employed in the
future... and we might even fail to envision them.

2) I'm myself in the camp of "MUST NOT use IEEE802 addresses or the
like"... but given that the requirements in this document have been kind
of wordsmithed among a number of groups and people, the more we tweak
the text, the more ths might take forever. I'd say that the current text
(if anything, modulo some minor edits) is good enough...

Just my 2 cents.

Thanks!

Best regards,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492