[AusNOG] IPocalypse eve?

[Matthew Moyle-Croft]

On 31/01/2011, at 12:47 PM, Damien Morris wrote:

On 31/01/11 12:26 PM, "Matthew Moyle-Croft" <mmc at internode.com.au<mailto:mmc at internode.com.au>> wrote:

They're mostly from people who don't appear to have done it.
We've got IPv6 dual stack for customers.  We sell ADSL CPE
which supports IPv6 (and more of our current ones will in the
next months).

What happens if you run dry of v4 for the dual stack? Is there a future
planned where customers will be v6-only, and how they'll reach v4 content
in that case?

We start to insert the evil layer called CGNAT.   This'll no doubt cover classes of customers who don't want to pay extra for their own IPv4 address.

(This is not indication of our particular strategy, just what'll likely happen generally.  Don't assume that having an unshared IP now won't mean you'll keep it in the future no matter how you misread someone's T&C).


The customers change their ADSL login from
username at internode.on.net<mailto:username at internode.on.net> <mailto:username at internode.on.net> to
username at ipv6.internode.on.net<mailto:username at ipv6.internode.on.net> <mailto:username at ipv6.internode.on.net>
on capable CPE and it just works. I'm tempted to get my 88 year old
grandfather (who bought himself a Win7 laptop recently) a dual stack CPE
just so I can say that it's that easy.
Most people have dual stack OSes these days so don't even have to do
anything.

I like that.. I'll try it once I have capable CPE. Have any particular
vendors/models proven themselves in the field to reliably dual-stack or
pure v6?

Billion 7800N and NL do dual stack.  Fritz!Box we're about to sell does it.  Netcomm have beta code for one model.  Billion are working on code for other devices.


Do the CPEs dole out world-routable v6 IPs to wired/wireless clients like
they currently do with RFC1918 space? Dual-stack public v6/private v4?

World routable v6.


e.g: /48s vs /56s vs /64s for consumers, /64s vs /126s for PtP, SLAAC vs
DHCPv6, ND vs DoS attacks, DNS64/NAT64 vs "it'll never work"..
DHCPv6 Prefix Delegation is important.   The subnet size is, well, not
very relevant, /64 is too small, but anything more is fine for almost
anyone.
DHCPv6 hopefully will be supported by more OSes and that'll make life
pretty
easy.

/64 is too small because the customer may want to subnet further and we
want to avoid NAT..?

Yes.   Not so much NAT, but v6 really needs /64 on a "LAN" segment.  /60 is what we're using on our trial and NOT one person has run into a problem with that.


Perhaps when it comes to IP it's possible to have too much of a good
thing? Or have most people worked it out and it's just the vocal minority
on nanog? :)
Too much noise, not enough doing as usual on NANOG.

:) I suspected, I just thought it would be a lot more settled by this
point. Is Internode's strategy fairly indicative of what we'll see from
the rest of the industry?

Dunno.   We have a strategy, ask the rest of the industry if they do.


Thanks for the reply MMC, I know a lot of this must be getting old to
discuss.

A bit.   Explaining DHCPv6 PD is getting annoying.  I think I might write some slides and just get on the conference circuit to tell people how this s**t works.

MMC


Cheers,
Damien.

_______________________________________________
AusNOG mailing list
AusNOG at lists.ausnog.net<mailto:AusNOG at lists.ausnog.net>
http://lists.ausnog.net/mailman/listinfo/ausnog

--
Matthew Moyle-Croft
Peering Manager and Team Lead - Commercial and DSLAMs
Internode /Agile
Level 5, 150 Grenfell Street, Adelaide, SA 5000 Australia
Email: mmc at internode.com.au<mailto:mmc at internode.com.au>    Web: http://www.on.net<http://www.on.net/>
Direct: +61-8-8228-2909      Mobile: +61-419-900-366
Reception: +61-8-8228-2999        Fax: +61-8-8235-6909

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ausnog.net/pipermail/ausnog/attachments/20110131/aa7313dd/attachment.html>