[Already__Taken]

Can IPv6 become, ironically, the reason ipv4 never dies? Once a majority move to v6 wouldn't that mean a whole bunch of the ipv4 space is being free'd up.

This allows those who never update to actually never update.

[ay]

Once the critical mass is reachable over IPv6, the rest becomes "long-tail" which will eventually be too expensive to bother about.

An extreme example: if an internet user has IPv6 and just uses gmail, youtube and facebook - they can turn off IPv4 right now and not notice anything.

Lee Howard has had an interesting presentation on the subject: https://www.nanog.org/sites/default/files/wed.general.howard...

p.s. Quite a few million of T-mobile's subscribers are also IPv6-only today, just that they use NAT64/464XLAT to connect to IPv4-only services.

[icebraining]

I think there will remain a minority, yes. Probably older devices which still work fine otherwise (especially appliances, sensors, etc). We'll probably see a bunch of such IPv4 LANs behind NAT64 gateways, so that the rest of the 'net can still talk to them.

But I don't think having more free space will make much difference; they can use a reserved IPv4 address (in the 10.* range or so) and have a public IPv6 address configured in the NAT gateway.

[yoha]

People who adopt IPv6 addresses tend to keep an IPv4 address as well.

[kalleboo]

Until we see more carriers moving users over to DS-Lite, where their IPv4 usage is NATed https://news.ycombinator.com/item?id=8680759

[netheril96]

But if a majority move to IPv6, wouldn't the holdouts have difficulty communicating?

[cm2187]

There has to be bridges to IPv4. There will be IPv4 only servers for a long long long time.

[Arnt]

Unlikely. It's difficult to remove the last v4-only device from a network. Much easier to just leave the migration at 90, 95 or 99% complete and keep the v4 addresses.