First, let me say that I work for a company that writes open-source networking software, and I'm also a gamer.
Next, let me remind you (hopefully, you already know this) that the RIR's are soon (the projected date keeps shifting forward and back) going to run out of IPv4 addresses they can dole out to ISP's, so the ISP's can grow their businesses. This isn't an end-of-the-world scenario! What will happen is there will come to exist a "secondary market" for IP addressing space, as people who are currently holding more addresses than they need find they can profit from delegating subsets of it to purchasers.
Many companies aren't happy with the idea of a secondary market for IP address space. They're looking for ways to mitigate their exposure to that business cost, to put it bluntly.
The end game is "Carrier Grade NAT." You know the NAT box you have at home? They're going to move it into the ISP's core network. You'll have an RFC1918 address (one of those 192.168...'s or 10...'s or etc), and you'll still have a home gateway. But its upstream side will be a tunnel terminated inside the ISP's NAT. This is called an "IPv4 Softwire." The only real upshot here is that your home gateway has an IPv6 address, and could start offering IPv6 access if customers began to demand it.
But let me bottom line Carrier Grade Nat for you: If you were to say that presently you are renting an IPv4 address from your ISP for a monthly fee (along with access), then in the future you will be renting port numbers. Not necessarily fixed port numbers (you will be sharing your IPv4 address with thousands of your neighbors, so you might not get the well-known VOIP port for your the fixed mapping you're paying for).
As gamers I think we all know how often we have to futz with NAT mapping for one game, voip service, or another. As gamers, will we be happy taking the NAT mappings our ISP is willing to afford us? I wouldn't be.
The alternative is deploying IPv6 to the home. More address space than you can imagine, and we're back to paying for straight access costs (not addressing or port space). This is simply not going to happen at the present rate. Why? Networks won't do it unless customers demand it. But look at all the websites, even of game companies, that don't have AAAA's on their systems (DNS records that resolve hostnames (as in URLs) to IPv6 addresses). There's simply no content available via IPv6, and this isn't changing. Why would any consumer want to buy a product they won't use? If the game companies don't see their own enlightened self interest in IPv6 adoption, why should a consumer see it?
The IPv6 pig won't fly without JATO assistance, it seems. Hence, my subject, there'll be a day, not long from now, that it will suddenly become a hot topic. The pity is that we could have talked about it for years beforehand.
So, I'm not here to start a debate. At least, not a heated one. I think the future is a hazy crystal ball, and I'm not sold that anyone has a good bead on it, myself included. But I thought I'd post something I could point back to later and say, you know, if people had taken a little more interest in the structure and design of the networks they were buying into (start demanding IPv6 access as consumers, start deriding games companies for not 'getting it', ...), maybe things could have worked out differently.
Some quick references I can remember -
[url]http://www.ietf.org/html.charters/softwire-charter.html[/url]
[url]http://www.nanog.org/meetings/nanog44/presentations/Tuesday/Durand_lightning.pdf[/url]
[url]http://www.ietf.org/proceedings/08nov/slides/softwire-5.pdf[/url]
