The 6bone is dead. Long live IPv6.
(Don’t know what IPv6 is?)
I’m no stranger to v6 - I played with it years ago while at 3Com (hi Cindy!) and actually deployed it natively, along with an automatic 6to4 tunnel relay, at the last ISP I worked at. However, I think I was on the wrong side of the problem. I thought by offering it to customers, I’d have users wanting it. But without content, why would they care?
IPv6 suffers from the “chicken and egg” problem (which is the surprisingly on-topic discussion on NANOG these days). Users don’t clamor for v6 because there isn’t any v6 specific content or content that they can’t already get. Content providers aren’t providing v6 content because users aren’t clamoring for it. But the technology to suppot it has been there for years - governments have mandated it and NASA’s even launched an IPv6 satellite.
But to quote Nathan Ward from NANOG,
“Because for IPv6 to be useful to the masses, content is required.”
And the masses are coming. Like Apple did with making 802.11b accessible in the iBooks and by making the Airport so affordable, Microsoft is doing to v6 in Vista (and to an extent in XP SP2) with Teredo (and don’t forget the mobile users on Verizon Wireless or Sprint PCS).
Three things push technology like nothing else: gaming, military conflict and porn. So while I might not agree with the content, the folks at The Great IPv6 Experiment are probably on the right path to spur consumer demand.
It’s my position that it’s up to providers to start providing and that includes Mozilla.
Over the past week I rolled out v6 connectivity to hosts within Mozilla’s San Jose colo (thanks to Sprint and ISC for ipv6 tunnels) and natively at Moco headquarters. In the near future, I hope to deploy v6 versions of Mozilla web properties. In fact, the Bugzilla folk are already drinking out of the v6 cup with their test server - landfill.bugzilla.org (or landfill.ip6.bugzilla.org if you want to force v6).
This is largely an experiment to try to understand what it means to support v6 users and v6 content in a v4 world. For instance, what happens to a user who has a global unicast v6 address but no global connectivity or user in the US with a v6 tunnel to someone like SixXS who tries to reach Mozilla properties and ends up with poor page load times?
As much as I’d like to, I probably can’t just flip the switch and enable v6 versions.
So my goals around this are:
- Deploy an open Teredo and 6to4 tunnel server hosted by Mozilla to give the community a doorway to ipv6. Microsoft maintains the default XP/Vista configured Teredo server but some may not want all of their v6 traffic going through Microsoft.
- Deploy v6 versions of Mozilla web properties.
- Understand end-user impact of living on a v6 host in a predominately v4 server world.
- Understand implications of deploying v6 versions of Mozilla web properties in a predominately v4 client world.
- Share results with the community.
Nathan has a good idea that he talks about here to figure out how many eyeballs content provders would lose by adding AAAA records. In the future I want to roll this out on some Mozilla web property. We get an amazing number of unique visitors each day that I’d have a good sample to test against and report on.
Some might see me as a kook but I don’t care . IPv6 is here and actually is the future whether we want to believe it or or not (and subnetting is so much easier). Long live IPv6!
ps. Firefox may not be configured to work with ipv6. In order to make Firefox play nicely with ipv6, you’ll need to make sure the following variable (in about:config) is set to false:
network.dns.disableIPv6