Re: implications of 6to4 for v6coex

[Rémi Denis-Courmont <rdenis@simphalempin.com>]

On Fri, 3 Oct 2008 00:01:04 -0400, "Erik Kline" <ekline@ekline.com> wrote:

>>>> Well it seems to be working well enough (or popular enough, anyway).

>>>> What if I said that 69.8% of IPv6-enabled users who visit google.com

>>>> have 6to4?

>>

>> That's the least I'd expect, as an author of 6to4 ;-)

>>

>> Any stats for Teredo?

> 

> This test was against a dual-stacked hostname.  Teredo comes in around

> 1.2%.  The same or less than ISATAP.



I expect that Teredo should be _forever_ close to 0% against *dual-stacked*

hostnames, thanks to source address selection. I have been proposing to fix

the address scoping rules of RFC3484 to get that number even lower: RFC1918

IPv4 addresses should global scope (which they are because of NATs), w.r.t.

to source address selection. It appears that Windows already does this; if

it did not it would prefer Teredo over native NAT'ed IPv4 in face of a

dual-stacked hostnames.



Teredo should be statically meaningful only with IPv6-only hostnames.



(...)

>>> I don't think that you can accurately say that that you tested all IPv6

>>> enabled users, because you aren't testing whether they are Vista and

>>> have Teredo enabled. Point em at a hostname with only AAAA and my

>>> expectation is that you'll significantly see different results - close

>>> to what I see, which are like:

>>> - ~90% Teredo

>>> - ~7-8% 6to4

>>> - rest Native

> 

> Yeah, but this doesn't represent reality for me as a content provider.



I tend to agree that content provider should not have to care much about

Teredo. To me, it seems more useful for peer-to-peer use cases. I would

say, that's a concern for "consumer ISPs", not for Google.



-- 

Rémi Denis-Courmont