On Sat, 02 Jul 2011 19:44:24 -0700
Doug Barton <dougb(_at_)dougbarton(_dot_)us> wrote:
On 07/02/2011 18:50, Mark Smith wrote:
Where is the evidence that 6to4 is holding back native IPv6
deployment?
It's been discussed ad nauseum in numerous fora.
Discussion isn't evidence, as people usually don't post any data to
support their assertions.
Since posting that question I remembered RFC6036 -
"Emerging Service Provider Scenarios for IPv6 Deployment". Here's what
is says about ISPs' views on 6to4. I don't think it supports the
assertion that 6to4 is holding back native deployment -
"2.5. IPv6 Technologies
Turning to technology choices, the overwhelming choice of approach
(94%) is a dual-stack routing backbone, and the reason given is
simplicity and cost. 39% run, or plan to run, a 6to4 relay as well,
and 16% run or plan a Teredo server. "
"Diffusions of Innovations" discusses attributes of innovations that
influence their adoption. One of those attributes is trialability,
meaning how hard or easy is it to trial the innovation. If a
innovation requires a high level of commitment to trial, then it is
less likely to be trialed and therefore adopted. Once an residential
ISP as IPv6 transit, 6to4 is a simple and much lower cost way of
trialing IPv6 services and gaining experience with IPv6, before and
during the development and deployment of native IPv6 services and the
required supporting infrastructure, such as AAA, helpdesk training etc.
Declaring 6to4 to be historic might encourage native IPv6 deployment,
but I think it will also make trialing IPv6 much harder.
Bad 6to4 (which almost
all of it is) results in a poor user experience when the largest content
providers enable AAAA records. Thus, they are less inclined to enable them.
The involvement in World IPv6 day by large content providers and the
apparent lack of significant problems would be suggest the opposite is
now the case. Google continuing to provide youtube video content to 6to4
tunnel users (such as myself), nearly a month later, suggests that any
problems with it are tractable. 6to4 users are easy to spot by the
2002::/16 prefix, so if Google needed to they could probably quite
easily limit their IPv6 content delivery to native only IPv6 users.
So, in other words, Google now consider 6to4 to be reliable enough to
continue to deliver content using it for one of their high profile
content sites.
I realize that there are a lot of people that dismiss both the evidence
that's been put forward and the rationale, but it's been presented and
discussed pretty thoroughly.
In Geoff Huston's report (Dec 2010), he says that he measured a 13%
failure rate with 6to4. A number of things have happened since then -
IANA has run out of IPv4 addresses, and World IPv6 day. Those two
events have increased the interest in IPv6 (the residential IPv6 trial
I was working on gained more participants due to these events),
which would have created an incentive for people to improve the quality
of their 6to4 operation, or possibly deploy a trial 6to4 relay. An
update on that data would be useful.
Regards,
Mark.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf