Re: [v6ops] draft-ietf-v6ops-mobile-device-profile last call- "harmfully broad"?

[Lorenzo Colitti <[e-mail]>]

On Mon, Feb 16, 2015 at 6:23 PM, Heatley, Nick <[e-mail]>
wrote:

>  Lorenzo,
>
> Focusing on a) what do you suggest?
>
> Have you another way?
>

Yes. Make IPv6* (see below) a requirement for carrier-branded devices, and
give the OEMs a credible signal that from date X onwards, you *will* fail
TA on every device that doesn't implement IPv6, and you *will not* waive
the requirement. That's what Verizon and T-Mobile did, and it worked for
them.

It's true that T-Mobile and Verizon have more market power than some of the
carriers that are behind this draft. However, they definitely do not have
more market power than all the carriers behind this draft taken together.
Also, when they put those requirements down, IPv6 support in mobile devices
was in a much worse state; some devices didn't implement it at all, and
some had pretty severe bugs. Today, any OEM that sells to T-Mobile or
Verizon must have an IPv6 stack that's high-enough quality that tens of
millions of users can use it. So the problem you need to solve is easier
than the one they needed to solve; all you need to do is ask them not to
disable IPv6 in the European firmware, just like it's not disabled in the
US firmware.

If your problem is that you want to go directly from IPv4-only to
IPv6-only, but you feel that you can't do that because iOS doesn't support
464xlat, then you're in the same boat as T-Mobile. And in fact, I believe
that their requirement is something along the lines of "if your device runs
an operating system that supports 464xlat, you MUST configure your device
as IPv6-only with 464xlat" (which is roughly equivalent to saying, "you
must do 464xlat unless you're Apple"). That has a lot of value, because
whatever percentage of your users are on Android or Windows Phone won't
consume IPv4 addresses. The public statistics say T-Mobile is close to 50%
devices running IPv6-only. That's a lot IPv4 addresses saved.

Over time, the more operators take this position, the more will be using
464xlat, and eventually even Apple will read the writing on the wall and
implement it. They're free to use the Android implementation; it's
Apache-licensed.

 The disappointed for me in the discussion about this common reqts paper is
> not that there are a few strong voices rejecting this initiative.
>
> It is that there are no voices of support. As I think you have pointed
> out, there is little support beyond the authors.
>

Personally, I think the resistance/lack of support is mostly due to the
fact that it is not the IETF's job to dictate/influence which technologies
are used in the Internet. The IETF's job is to specify those technologies,
but that's already been done.