Yesterday's Windows update causes IPv4 to be default

Doug Barton dougb at
Wed Nov 21 19:03:42 CET 2012

On 11/21/2012 4:02 AM, Phil Mayers wrote:
> On 21/11/12 11:58, Nick Hilliard wrote:
>> On 21/11/2012 01:54, Doug Barton wrote:
>>> Oh c'mon, that's just silly. The intermittent problems are almost
>>> certainly
>>> not the fault of the content end of the system, they are almost
>>> certainly
>>> on the user end, or somewhere in between.
>> I don't know about that.  Anecdotally, I've noticed that my ipv6
>> transport
>> causes fewer problems than random web sites doing stupid things.  E.g.
>> losing ipv6 connectivity for a week last year is one
>> which
>> comes to mind.
> This matches our observations. The (few) problems we've seen have been
> services which advertise AAAAs but don't function correctly.
> Handy hint: if you're running ISC bind as a nameserver and have a
> version with RPZ, you can selectively transform such AAAAs into NODATA
> responses for your clients, as a temporary workaround.

I'm responding pseudo-randomly to Phil here but Nick and Lorenzo are 
making the same mistake. :)

It's great that we all have some experience in the early days of 
deployment, and that the experience we have at this point is mostly 
good. But we haven't even scratched the surface yet. Less than 1% of all 
Internet traffic is over IPv6 right now (Google's recent announcement 
aside). Sure, most of the big sites have it mostly enabled, but the 
majority of content sites (numerically, not by traffic) don't. Same with 
ISPs, arguably by both metrics.

We have a LONG, LONG way to go before we will have anything approaching 
even a solid minority of traffic over IPv6, and there will be a lot of 
bumps along the way. Assuming that because the problem is mostly fixed 
at this point in time, for us (who do not represent the average Internet 
user), and therefore we don't have to worry about IPv6 vs. IPv4 
reliability problems anymore, is pure folly.


More information about the ipv6-ops mailing list