On 11/26/2013 12:31 AM, Gert Doering wrote:
Hi,
On Tue, Nov 26, 2013 at 05:22:15PM +0900, Lorenzo Colitti wrote:
Are you suggesting that we should have designed 464xlat on day one instead
of DNS64? That's a bit like saying that we should have designed fiber
optics without having copper. If DNS64 had not been designed and
implemented we wouldn't have 464xlat today.
I think he's saying that everyone should be using dual-stack, because
that's so much easier to roll out and maintain, and there's still plenty
of IPv4 left in the US region.
(Let me say that if *I* had a mobile network where I pretty much know
what sort of devices and stacks are being used, I'd go for IPv6-only PDP
plus 464xlat as well. On "DSL style" end user connections, with a
gazillion of unknown devices and applications, I'd more likely go for
IPv6+DS-Lite or IPv6+MAP - or full dual-stack *if* I had the addresses,
which is one of the small obstacles Doug seems to conveniently overlook...)
Which uses more IPv4 addresses, a traditional IPv4 NAT or 464xlat? At
the end of the day the PLAT still has to talk to the v4 net.
And frankly I take offense at the gratuitous American bashing here. I
worked my ass off for years publicly and privately pushing v6 adoption
and readiness; before, during, and after my time at ICANN. I continue to
work hard to encourage my customers and others privately, in spite of my
ever-increasing distaste at the mess the IPv6 literati have made of
deployment generally and the protocol specifically. And I've done all
this precisely BECAUSE I could do the math on the IPv4 runout, and I
still believe the Internet is too important a thing to be left
exclusively in the hands of the privileged few.
When I argue against ivory tower purity and in favor of practical
solutions that meet real world needs it's not because I'm trying to make
points on some imaginary scoreboard. It's precisely because we needed to
deploy IPv6 widely 10 years ago, and every misguided path we send people
down, every pointless bell or whistle we add to the protocol, and every
time we say "no" to completely reasonable requests like DHCP parity we
delay that widespread deployment. And that pisses me off.
So you should feel free to make any technological argument you want, and
if you can point out an area where I'm demonstrably wrong I'll thank you
for it. But don't insult me personally because I dare to disagree with
your opinion.
Doug
