Re: [v6ops] Please review the No IPv4 draft

[Matthew Petach <[e-mail]>]

On Thu, Apr 17, 2014 at 7:48 AM, Ted Lemon <[e-mail]> wrote:

> On Apr 17, 2014, at 6:56 AM, Owen DeLong <[e-mail]> wrote:
> > Providing a mechanism for an IPv6 LAN to shut down the IPv4 mobile
> network on a device which is unlikely to be enterprise controlled is not a
> solution in the real world, but only an attack vector.
>
> We are in full agreement on this point.   It has to be up to the device
> what it does with whatever configuration information the network supplies,
> and *I would not expect a device to take advice from one network about how
> to operate on another.*   You could imagine exceptions to this, but I think
> all of them would be equally well addressed by individually updating each
> device, so there's no reason to have a protocol for doing it.
>

Ted, I just want to repeat your statement once more:

"I would not expect a device to take advice from one network about how to
operate on another."

I don't think I could sum up my objections to this draft any
more clearly than you just did.  I would not expect any device
of mine to take advice from an IPv6 network about how to
operate on an IPv4 network.  Even if those networks happen
to be sharing the same physical piece of wire, or same range of
RF spectrum, they are independent and orthogonal networks.

Now, one can take steps to bring those networks together,
such as creating tunnels, encapsulating one network within
the other; but under those circumstances, one should realize
the now-shared-fate scenario one has delved into.  But by
default, the two protocols were designed to be ships in the
night, and what is being proposed here clearly violates
that.

(looking at this from the other side, it would have been an
interesting proposal a few years back to have a DHCPv4
option to tell hosts to shut down all those pesky teredo,
6in4, and other pesky IPv6 tunnelling, translation, and
encapsulation technologies that the hosts might have
tried to use that caused extra unnecessary packets
to clog the networks.  I'd like to imagine the outcry
against such a "packet of IPv6 death" would have
been equally as loud as what you're seeing now.)

Matt