Google and IPv6

Remi Denis-Courmont rdenis at simphalempin.com
Mon Mar 17 13:10:36 CET 2008


On Sun, 16 Mar 2008 19:34:56 -0700, "Erik Kline" <ek at google.com> wrote:
>>  I agree with what other posters have mentioned about AAAA records not
>>  being ready for primetime on a service like www.google.com :) Potential
>>  reachability issues are nothing to be trivialized and in my limited
>>  experience quad-A records have had a somewhat unpredictable effect upon
>>  DNS resolution time / success rate / complexity.
> 
> Speaking personally, I have seen analysis of neither client behaviour
> nor connectivity in the IPv6 Internet today.  Are things really as bad
> as folks say or is it partly a kind of "urban legend of horribleness"
> that persists from earlier tests with less-mature operating systems
> and less reliable connectivity?  I  just have 6to4 at home and my Mac,
> Linux, and XP boxes all seem to work just fine.

6to4 at home is one thing...

I have seen my 6to4 setup fail. For instance, some hotels do assign public
IPv4 addresses through DHCP, but yet they blackhole proto-41. Stateful
firewalls also customarily drop proto-41 toward the native IPv6 Internet,
because packets from the downstream 6to4 relay come from an "unsolicited"
IPv4 address.

As far as I know, XP nor Vista run any kind of connectivity checks before
bringing up 6to4 automagically, so...

Also, you do want to maintain your own 6to4 and Teredo relays, and I really
mean _maintain_, not just setup-and-forget. Otherwise, you may end up
routing through congested, if not dead, relays anywhere else. I have seen
this happen - in fact, it's happening right now with my the collocation
company hosting my personal server :(


Because of these issues, I have been suggesting that RFC3484 be modified to
favor IPv4, including private, addresses over Teredo, and even over 6to4.
But it seems most people disagree, or don't care.

-- 
Rémi Denis-Courmont
http://www.remlab.net



More information about the ipv6-ops mailing list