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

[David Farmer <[e-mail]>]

On Thu, Aug 17, 2017 at 3:48 PM, Brian E Carpenter <
[e-mail]> wrote:

> On 17/08/2017 21:05, DY Kim wrote:
> > ---
> > DY
> >
> >
> >> On 17 Aug 2017, at 17:58, Ole Troan <[e-mail]> wrote:
> >>
> >> Have you read RFC7421?
> >
> > Yes, I did. With my own bias, the doc looks quite in the interest of
> defending one school of thoughts.
>
> I am puzzled by that remark. The authors included a range of opinions and
> we worked
> hard to presents facts (both facts about the published specifications and
> facts
> about observed behaviour of implementations). And the opening paragraph
> states
> "IPv6 routing is entirely based on variable length prefixes (also known as
> variable
> length subnet masks), there is no basic architectural assumption that
> n has any particular fixed value."
>
> So I'm afraid that the document is actually stating something about
> reality,
> whether we like it or not.
>
>     Brian
>

I think RFC7421 fairly represents the facts as they are, in section 4.3.2
"Other Observations", it even says;

   Also, DHCPv6 is in widespread use without any dependency on the /64
   boundary.  Reportedly, there are deployments of /120 subnets
   configured using DHCPv6.

I just wish we could get RFC4291bis to represent the facts as fairly as
RFC7421. No, that doesn't mean that SLAAC with anything other than /64 and
64 bit IIDs should be allowed by RFC4291bis. However, it seems clear that
other than for SLAAC, /64 is just a recommendation, an important
recommendation all be it, but a recommendation and not a requirement. /64
is a requirement for SLAAC, changing that is out of scope for RFC4291bis as
I see it.

Personally, I'm just looking for RFC4291bis to represent the reality of
IPv6 today.

My personal version of that reality is; at work and at home my iPhone,
iPad, and MacBook Pro all get IPv6 addresses with SLAAC from a /64 subnet,
and I'm happy with that.  I wish Comcast would give me more than one /64 at
home. Maybe they do now, I haven't had time to screw around with my home
network in 12 to 18 months.  However, I've had native IPv6 at home from
them for 3.5 years now, so I'm doing better than a lot of the world. My
iPhone and iPad haver several IPv6 addresses from my carrier AT&T, but I
don't seem to use them for Internet access.

At work we are #89 and at 63% of our hosts doing IPv6 for July on
http://www.worldipv6launch.org/measurements/ and we do SLAAC for most
things.

However, my MacBook Pro does allow for manual configuration of an IPv6
address and an on-link prefix of any length, and while I haven't tested I
suspect the DHCP client will work for with other than /64 too.  Also, most
laptops and desktops, and all our routes at work allow manual configuration
of any length subnets and work with DHCP with any length subnets as well.

So can we please get past arguing about letting SLAAC use subnets other
than /64, and update RFC4291bis to reflect the reality of IPv6 fully
discussed in RFC7421. Once we've done that, then maybe we can look at
retooling SLAAC, but that is not simply changing a few words in RFC4291bis,
that is a major redevelopment effort, with new running code, and testing
like crazy. That will be a lot of work, it may or may not be worth it, but
that should come after we get RFC4291bis updated to the real world
described in RFC7421 out the door.

Thanks.

-- 
===============================================
David Farmer               Email:[e-mail]
Networking & Telecommunication Services
Office of Information Technology
University of Minnesota
2218 University Ave SE        Phone: [phone]
Minneapolis, MN 55414-3029   Cell: 612-812-9952
===============================================