Perhaps too late, it's widely recognized now that the IPv4 address pool will run out in the next few years. The current prediction, based on current data, is for the IANA to exhaust its pool on Jan. 25, 2011, and for the last block from the last RIR (Regional Internet Registry) to go on Dec. 26, 2011. Perhaps someone will buy it in a Boxing Day sale.
The answer to the problem has been around for ages: IPv6. It's supported by all major operating systems and network hardware vendors (at least for enterprise equipment), and many tests have shown it to be practical. Problem is, it's incompatible with IPv4, so all networking software needs to be updated to support it.
A number of proposals have been made for a transitional software setup to help people move from IPv4 to IPv6 (Comcast made one just the other day), but it's clear that these are moving users too slowly, if at all. We'll run out before a lot of people have moved to IPv6.
Of course, when we say we've "run out" we mean the official registries have run out of them. The addresses all will still exist, in the hands of other people. One solution is to allow a market to develop in these addresses so that they can move to those who need them based on supply and demand. Currently this is not permitted; you can't sell the address blog that your RIR allocated you.
This is especially a shame, since addresses were allocated in large chunks with no real concern back in the '80s and even '90s, and as a result there are many organizations with huge blocks of addresses they don't need, but they have no reason to give them up.
These are some of the problems considered by the Internet Governance Project, a consortium of academics with Internet expertise. They recently issued a report titled "Scarcity in IP Addresses: IPv4 Address Transfer Markets and the Regional Internet Address Registries," which considers several approaches to the problems.
The report is enthusiastic about a market as the solution to the problem. It calls a market a "pragmatic solution" to the problem of optimizing use of the limited resource of IPv4 addresses. It says that the risks of a market approach are small compared with the potential benefits and that the change is not really so radical. On the other hand, failing to create a legitimate market runs the definite risk of creating an illegitimate one, which would create significant problems through inaccurate registrations.