On Jul 3, 2011, at 12:02 AM, Cameron Byrne wrote:
On Sat, Jul 2, 2011 at 8:10 PM, Keith Moore
<moore(_at_)network-heretics(_dot_)com> wrote:
On Jul 2, 2011, at 3:21 PM, Cameron Byrne wrote:
I saw the same thing. It is a shame that work that directly removes barriers
to REAL ipv6 deployment gets shouted down by a few people not involved in
REAL ipv6 deployment
I find myself wondering what you mean by REAL IPv6. For me, REAL IPv6 is
code that uses the IPv6 programming model, 128 bit addresses, end-to-end
transparency, no NATs. 6to4 certainly qualifies.
That's not what it means to me. REAL IPv6 is a replacement for IPv4
and can address greater than 100s of billions of endpoint and is
suitable for very large traffic loads. As an access network provider,
i need content on native IPv6. It does not make sense to anyone in my
organization or industry to deploy IPv6 unilaterally. There is no
benefit in this approach vs just doing NAT444. If there is IPv6
content on a meaningful scale ( by the numbers that means for "my
network": Google, Facebook, Yahoo and their CDNs ...), then i have a
solid business case for IPv6 access networks. Full Stop.
Chicken-and-egg. You can't justify widespread deployment of IPv6 until there
are a lot of content/people/applications using it, and there won't be a lot of
content/people/applications using it until it's widely available.
The whole purpose of mechanisms like 6to4 is to help break that logjam. We
need to fix the existing mechanisms, and/or provide better ones, rather than
killing things that work... even if they only work in corner cases.
(Basing the argument entirely on "content", IMO, is misguided, because it
neglects significant potential drivers of IPv6 adoption, and because there's
still no incentive to move to v6 as long as the same content is available on
v4.)
If the content guys say 6to4 is a pain, and they do, then i need to
help them find a way to solve that pain.
So you want to help the "content" guys at the expense of other legitimate uses
of 6to4. Frankly, I don't think that's reasonable. The net doesn't exist
just for content delivery.
In the meantime, i null route the 6to4 anycast address because it
creates half open state in my CGN. Been doing that for at least 5
years. My next step is filtering AAAA over IPv4 access because 6to4
client brokeness won't die on its own, that will be rolled out in a
few months.
So you admit you're sabotaging the network?
Keith
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf