True. But other than outright regulation / fines... I'm not sure how to hit the right people. Right now we have:

a) ISPs being sluggish
b) Some network people at companies being obstinate
c) Companies being irresponsible about their own conversion
d) The government not leading the effort (though in all fairness the Obama administration is better than I would expect on IPv6 issues).

Makes sense. This is where I wish ARIN were pushing much harder to break the chicken and egg issue.

I hope ARIN doesn't reclaim. Let the shortages create the pain.

That's true. One big advantage of IPv6 wii be the ability to create an essentially infinite number of external IPs associated with VMs / containers / microservices.

They did that. ARIN didn't want to raise prices they just blew through their IPv4s. So ISPs will need to buy them from each other rather while IPv6 addresses will given out for free.

The first step is the carriers / ISPs getting everyone an IPv6 address. The first thing to break after that will be geolocation as the carriers start pooling their home / small business IPv4 addresses and allocating them from a single common pool (so all Verizon originates in West Virginia). That will give companies a reason to switch their consumer internet.

In terms of B2B... I suspect most companies will change most stuff. However longer terms routing tables are getting too fragmented for some many routers and there are overlapping addresses (i.e. we don't have 1 address goes to one place anymore, especially in the 3rd world). As that gets worse IPv4 will break and companies will change.

The big question is: why haven't the telcos moved home / small business over yet?

Why do you need multiple IPs on a machine with 1 nic? Besides the last 64 is going to be the MAC address.

AFAIk the 2001:678::/29 provider independent space are just supposed to be a bunch of /48 relays.

For residential users on IPv6 they can likely now have a fixed IP (or a /60) now. So it would only change when they their home ISP.

A local internet registry at smallest only gets a /32. Which means they can only do 65k homes which would be too small. Now admittedly if they are getting a /20 (maximum allocation) will be fine. But that means you can only have 256 locals in each regional (i.e. each /12) and that's likely too few. There is tons of room but it isn't infinite.

    I'm a fan of a /60 for homes. I guess you are right there is enough room to make a /48 work but that seems like needlessly throwing away a lot of bits.

You really shouldn't be using NAT with IPv6. The idea is one machine, one address. Given DNS I'm not sure why you would want fixed fully external IPs. But if you do you want to have some fixed external addresses do it via. some sort of relay where you have an external server at a telco colo with a very long term address and then the telco wires it back to your server (or just host your server with the telco). I think part of the idea of IPv6 is simplifying the routing tables so the old any address can go to any physical location should die. Routers should as much as possible be making routing choices not based on lookups buy literally picking one or more address bits and assigning them to a physical wire, very low latency routing.

You sure that's residential? Why would a home need 65k subnets?

Logical fallacy: appeal to popularity. But I'm willing to bet I've got orders of magnitude more Score: 5 comments than you do in any case.

Obviously you were just trying to be clever, but this is a lot of bollocks. Even many machines which don't have any PCI bus connectors have a PCI bus...

Remember this is the internets, and if you can't use google, you're gonna have a bad time.

https://www.blackhat.com/us-15...

https://github.com/jbangert/tr...

I searched "flaws in intel mmu" and got these results back in the top ten. Perhaps you should learn to internet, coward.