Eventually in the future, people will come to realize the over-exaggeration of running out of ipv4 addresses. The unused blocks will come out of woodwork from everywhere. Well, I hope my opinion is wrong.
There's still a finite number of IP addresses that is below all reasonable expectations for how many devices will be wanting an IP address. Reclaiming unused blocks just softens the shock by making the exhaustion less sudden and more gradual. We're still definitely going to run out.
There are a little over 4 billion IPv4 addresses. If they cost even 10c / month the vast majority of them would still be unused. However, because they where free well over half of them are simply wasted and we are "running out".
Assuming you mean the American national debt, not really. That's only $480 billion/year. That wouldn't even plug our current deficit, let alone start paying off the debt.
By future do you mean right now where NAT is being used by some providers already? I know Rogers in Canada has all mobile devices going through NAT unless you pay a nice "VPN" fee.
Also the assigned but unused IPv4 blocks don't do anything for other providers that actually need addresses.
Rogers only blocks PPTP VPN connections. Also it's not a VPN "fee" exactly, it's an entirely different APN. If your device doesn't allow you to manually edit the APN, you have no hope of getting your Rogers device to create a PPTP VPN connection over 3G because you cant access the internet.vpn.com APN.
However, L2TP and IPSec VPN connections work just fine over 3G on Rogers, no fees or APN changes required. So I'm not positive that NAT is the issue if I can establish the other types of VPN connections although this is hardly my area of expertise.
Once no more unused addresses turn up, they can start economizing on the used ones too. For example my university gives a publicly routed IP address to EVERY computer, even though most are completely firewalled off.
IBM has an entire /8 that they use internally for addressing their computers, none of them are publicly routable or even publicly accessible.
Even if they were to start announcing them or selling them it would be a HUGE project for IBM internally to make sure that their computers aren't going out on the web for internal resources, and it becomes an huge issue of having to renumber.
Unfortunately there is no good way to support something like that. Some companies didn't need to use the private LAN addresses because they had their own /8 and changing it now would require too much effort and work.
They wouldn't have to fix the whole thing.
They could carve out /16 blocks, one at a time and give them back, starting with the ones that are currently not in use.
I see that it would be a lot of work, but so will switching to IPv6 be. The question is: which will be worse? which will actually happen? I don't know.
I'm probably forgetting a lot of things but to me it would seem doable to switch to private LAN addresses using DHCP and NAT, as long as there are no hardcoded IPs lingering around (but only hostnames which can be updated in a central location).
Very valid point - especialy given that many text-books and other Ip education material would tend to bias you to use 10.0.0.0/8 as that is what you do, remember many companies who used the old SCO IP range as there `network chap` learned IP from a popular book that used SCO IP's in there examples.
Thing is though, if a pretty compitent company like IBM who know how to do IT have not moved to IPV6 and use public IP ranges for there internal networks albiet router blocked. Well, it just don't bode well for other less technicaly compitent people to move now does it and the cycle of procrastination with IPV4 carry's on :(.
And this is exactly how it is supposed to be! The solution is switching to IPv6, not more NAT and ever more complicated routing to support the fragmentation of networks.