[dhcwg] FW: I-D Action: draft-templin-duid-ipv6-01.txt

[Antonio Escalera <[e-mail]>]

Hi Fred,

I responded inline below.



On Mon, Jan 11, 2021, 2:45 PM Templin (US), Fred L <
[e-mail]> wrote:

> Hi Antonio,
>
>
>
> Continuing on from the text you quoted, RFC8415 goes on to say: “Clients
> and servers
>
> MUST NOT restrict DUIDs to the types defined in this document, as
> additional DUID types
>
> may be defined in the future.” And, this document is exactly such a future
> document.
>
This quote is referencing DUID types, not DUIDs themselves. Additional
types are permitted, but interpreting meaning from a DUID in this manner is
explicitly prohibited in this case. In fact, it is the express intention of
the authors of (RFC8415)[1] that the DUID be irrespective of physical or
logical changes to the configuration of the networked device:

"That is, the DUID used by a client or server SHOULD NOT change over time
if at all possible; for example, a device's DUID should not change as a
result of a change in the device's network hardware or changes to virtual
interfaces"

I would suggest submitting a document to supersede RFC8415 first. I don't
understand your reasoning for wanting to change these requirements --
perhaps you have valid reasons. Nonetheless, a new DHCPv6 proposal would be
required before this can be considered.

>
>
> Also, if the text you quoted were taken to the absolute letter of the law,
> then there
>
> would only be one DUID type for all time, with only the length field
> differentiating.
>
False. To continue quoting (RFC8415)[1]


DHCP servers use DUIDs to
identify clients for the selection of configuration parameters and in
the association of IAs with clients.  DHCP clients use DUIDs to
identify a server in messages where a server needs to be identified.



IA, or Identity Association, is defined in the same document (here)[2] as:




The goal of DUID is to associate IA's with a client. Each server and
client has only one.

Your proposed RFC would prohibit multiple interfaces.



But, we already have four DUID types (DUID-LL, DUID-EN, DUID-LLT, DUID-UUID)
>
> and this document would be introducing a 5th type. So, somewhere along
> the line,
>
> someone decided that it would be useful to distinguish different DUID
> types.
>
>
>
The different DUID types in existence do not violate the aforementioned
quote in the previous email.

> Thanks - Fred
>
>
>

[1]: https://tools.ietf.org/html/rfc8415#section-11

[2]: https://tools.ietf.org/html/rfc8415#section-4.2


Conflicting standards with no clear reasoning for divergence and no
documentation defining changes are useless.

Regards,

Antonio

>