I can see it’s time to present a NANOG talk “How IPv6 saved the Internet” I’m just going to turn this thread into slides. The talk writes itself! -mel via cell
On Apr 1, 2025, at 5:03 PM, Mark Andrews via NANOG <nanog@lists.nanog.org> wrote:
Single IPv6 stack in the home is achievable today. Single IPv6 stack in the enterprise is achievable today. Single IPv6 stack in the access network is achievable today. This is how most phones work today. The node if it needs to talk to an IPv4 server uses one of the forms of IPv4AAS. Most probably 464XLAT. There are billions of devices that do that today.
Today if you attend conferences your IPv4 address is most probably 192.0.0.1 if they support IPv6-mostly (IPv4 option 108) and you have an up to date OS.
Enabling IPv6-mostly on the access network will allow your IPv4 lease load to drop as CPE start to support it.
The hard part will be turning down the IPv4 connections to your peers when the rest of the world has switched to IPv6.
Mark
On 2 Apr 2025, at 10:05, Alex Buie via NANOG <nanog@lists.nanog.org> wrote:
At the risk of oversimplifying:
“just because you can divide 8 apples between 2 people with long division on paper, it’s a heck of a lot easier just to count off if you can get away with it”
But this raises an interesting question; is there ever an actual end? What does IPv6 adoption completion look like? Is there a definition of success? You can almost certainly not buy a physical GP computer today at least that does not support static IPv6 addressing in some manner, at least in terms of network elements. And I could get IPv6 connectivity at nearly any address in the US, (where I have experience) if I’m willing to pay enough money for it. what else is there? Is IPv6 only considered successful with IPv4 is truly turned off? There’s always gonna be people who want 32 bits instead of 128 because it’s easier to carry around in your brain and doubly so if you don’t even need the rest of those bits for what you’re trying to do.
The allure of only having to set up a single protocol stack is very strong, but I just don’t see it happening in my lifetime in a “production” capacity.
You’re always gonna at least need some broker box somewhere with your last /29 talking to the vestiges of the companies still running exchange 2003. _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/MDVNX6WM...
-- Mark Andrews, ISC 1 Seymour St., Dundas Valley, NSW 2117, Australia PHONE: +61 2 9871 4742 INTERNET: marka@isc.org
_______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/7BQSH35W...