Re: [v6ops] I-D Action: draft-ietf-v6ops-unique-ipv6-prefix-per-host-07.txt

[神明達哉 <[e-mail]>]

At Fri, 18 Aug 2017 21:46:18 +0200,
Alexandre Petrescu <[e-mail]> wrote:

> Here, still it makes sense to allow for prefixes shorter than 64 (i.e.
> unique /63 prefix per Host, such that it can further grow the network by
> becoming a Router).
>
> The draft says:
> > a Unique IPv6 prefix (currently a /64 prefix) and some flags.
>
> That "currently a /64" sounds as a hardcoded value.
>
> It should be something like: "a variable length whose value could be /56
> for example; it could also be /64".

Today you can't use a prefix for SLAAC unless its length is 64 due to
the facts that
- All today's defined IIDs have 64 bits in length
- Hosts reject the prefix unless the sum of the prefix's length and
  the IID length equals to 128 (which means the prefix length must be
  64) according to RFC4862

You may not like it or might want to say "it doesn't have to be that
way"), but that's the fact today.  As long as
draft-ietf-v6ops-unique-ipv6-prefix-per-host talks about today's
practice, I don't think anything that doesn't work today is in the
scope of this document.

We may or may not choose to include the specific number of 64 in the
doc, on which I don't have a strong opinion.  But ""a variable length
whose value could be /56 for example" is very misleading since /56 for
SLAAC doesn't work today (perhaps "could" tries to express the subtle
nuance, but that would be quite misleading).

--
JINMEI, Tatuya