Re: [v6ops] Chair decision on WGLC for draft-ietf-v6ops-dhcp-pd-per-device-04

> section 8 explains

It is exactly the problem. Section 8 explains nothing, it demands and misleads:

“all IPv6 hosts are required to support SLAAC [RFC8504<https://www.rfc-editor.org/info/rfc8504>]”
“Choosing longer prefixes would require the client and any connected system to use some other form of address assignment, which many hosts do not support, and therefore limits the applicability of the proposed solution.” (it is a wrong statement by the way).
“prefix length of /64 or shorter is required to extend the network using [RFC7084<https://www.rfc-editor.org/info/rfc7084>] (see requirement L-2)”

All statements are misleading because taken out of the context of some SLAAC discussions (where it was useful).

Put in section 8 the clear statement that “for DHCP of local assignments /64 is not mandatory, prefix could be any size”.

>FWIW, if what you are suggesting is something like, "host gets a /96 (or /114 or /120), sends an RA to other devices, and uses DHCPv6 >to assign addresses to hosts", then RFC 7421 says:

>   using an IID shorter than 64
>   bits and a subnet prefix longer than 64 bits is outside the current
>   IPv6 specifications, so results may vary.

I am not lazy to check. It is a wrong reference because it was specifically for the SLAAC case: “the design of SLAAC, using an IID shorter than 64 …” The authors did not mean that it is relevant for DHCP too.

From: Lorenzo Colitti [mailto:lorenzo=40google.com@dmarc.ietf.org]
Sent: Thursday, November 2, 2023 10:31 AM
To: Vasilenko Eduard <vasilenko.eduard@huawei.com>
Cc: Lorenzo Colitti <lorenzo@google.com>; Brian E Carpenter <brian.e.carpenter@gmail.com>; v6ops@ietf.org
Subject: Re: [v6ops] Chair decision on WGLC for draft-ietf-v6ops-dhcp-pd-per-device-04

The draft already says that SLAAC does not need to be used:

   Note that assigning a prefix of sufficient size to support SLAAC does
   not require that subtended nodes use SLAAC; they can use other
   address assignment mechanisms as well.

As for prefix size: section 8 explains why the draft recommends that the network provide SLAAC-sized prefixes. Could the network provide longer prefixes and have stuff work? Maybe, if only the host needs addresses. But there are lots of use cases that won't work - RFC 7084, SNAC, non-DHCPv6-capable devices, devices that support DHCPv6 but not for all functions, etc.

FWIW, if what you are suggesting is something like, "host gets a /96 (or /114 or /120), sends an RA to other devices, and uses DHCPv6 to assign addresses to hosts", then RFC 7421 says:

   using an IID shorter than 64
   bits and a subnet prefix longer than 64 bits is outside the current
   IPv6 specifications, so results may vary.

This is not something I think we should recommend in an operational document.

On Thu, Nov 2, 2023 at 7:17 AM Vasilenko Eduard <vasilenko.eduard=40huawei.com@dmarc.ietf.org<mailto:40huawei.com@dmarc.ietf.org>> wrote:
Hi Lorengo,
Please, add to the draft that “DHCPv6 is possible between the proxy/client and the final host then /64 is not mandatory, prefix may be smaller.”
Eduard
From: v6ops [mailto:v6ops-bounces@ietf.org<mailto:v6ops-bounces@ietf.org>] On Behalf Of Lorenzo Colitti
Sent: Wednesday, November 1, 2023 10:37 PM
To: Brian E Carpenter <brian.e.carpenter@gmail.com<mailto:brian.e.carpenter@gmail.com>>
Cc: v6ops@ietf.org<mailto:v6ops@ietf.org>
Subject: Re: [v6ops] Chair decision on WGLC for draft-ietf-v6ops-dhcp-pd-per-device-04

On Wed, Nov 1, 2023 at 7:03 PM Brian E Carpenter <brian.e.carpenter@gmail.com<mailto:brian.e.carpenter@gmail.com>> wrote:
> The WG needs to steer clear of hard-coding /64 into drafts because it's
> causing problems with getting consensus on IDs.

That's 100% correct and although I remain supportive of the draft, it certainly
needs to be reworded to avoid mentioning /64 as anything but the present
de facto boundary. We still have that bug in the addressing architecture
standard too.

Brian,

Can you explain what you think should be reworded? I think the draft only uses "/64" in examples and figures, and where required by other documents such as RFC 7084 and the SNAC drafts (which it obviously cannot change). In actual text it's careful to say things like "a prefix of sufficient size to use SLAAC". It does mention that at the time of writing the only size is 64 bits, but that seems like a useful statement to make.

Cheers,
Lorenzo
_______________________________________________
v6ops mailing list
v6ops@ietf.org<mailto:v6ops@ietf.org>
https://www.ietf.org/mailman/listinfo/v6ops