Yesterday's Windows update causes IPv4 to be default
lorenzo at google.com
Wed Nov 21 04:01:22 CET 2012
On Wed, Nov 21, 2012 at 11:10 AM, Doug Barton <dougb at dougbarton.us> wrote:
> ... and with this perspective I sympathize, but I think the reality is
> that we are going to have wacky IPv6 connectivity problems well into the
> next decade, during the long ramp-up of knowledge and experience on both
> sides of the network.
Our numbers disagree. For users of MSIE on Windows Vista/7 (i.e., a
combination with no happy eyeballs at all), aggregate reliability to Google
dual-stack sites is over 99.95% of aggregate reliability to Google
IPv4-only sites. If you exclude Japan, which has its own special issues,
and another handful of networks with chronic IPv6 problems that can be
addressed using tactical fixes, that number goes even higher, coming pretty
close to 99.99%. (Even though it's just started rolling out, the beneficial
effect of the new Windows 7 on those numbers is substantial.)
I would argue that that kind of performance is acceptable for the vast
majority of applications, especially given that random outages/failures are
likely to lower IPv4-only reliability substantially below 99.95% anyway.
Yes, happy eyeballs can give you even more reliability, but it comes as a
substantial cost in complexity, and given the reliability numbers we have
today, it's not clear that it's worth it.
We are also going to continue to see reluctance on both sides if the
> hosts/apps are not robust enough to handle said wacky networks without
> significant degradation to the user experience.
The current level of IPv6 reliability has not stopped 5 of the top 10
global websites from deploying IPv6, nor has it stopped >60 major networks
rolling out IPv6 to substantial percentages of their users (see the World
IPv6 Launch measurements).
I would argue that as IPv6 growth continues, its importance will increase,
and bugs will be fixed. And the best way to ensure bugs are fixed is to
ensure someone notices - which is what Microsoft is doing.
Fear of IPv6 brokenness was a big reason for lack of IPv6 deployment for a
few years, but I think we're past that now. The concerns are back to what
they were before - no perceived business case, little demand from users,
lack of support in equipment, and not a priority for the organization, and
so on. And that is changing as IPv6 rolls out more widely and IPv4 address
space comes under more and more pressure.
So I don't think we need to concentrate our efforts on this problem now, we
just need to wait for operators to deploy it; the bugs will work themselves
out. Of course it will take time, the Internet is a big network. But
Google's public IPv6 numbers hit 1% this weekend, and growth was 2.94x year
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the ipv6-ops