Psst! Wanna Buy A Black-Market IP Address?

Some problems always seem so far away -- until you turn around and find them chomping on the seat of your pants.For many years now, IT experts have known that the world's supply of IP addresses was inadequate. This trend, known as "IP exhaustion," is the main driver behind the move from IPv4 -- the prevalent version of the Internet Protocol standard -- to IPv6.

This transition will solve the problem of IP address scarcity -- if not forever, then certainly for generations to come. While IPv4 provides for a little over four billion unique IP addresses (in theory -- the actual number is smaller in practice), IPv6 will up the ante considerably.

How considerably? We're talking about a number with 38 zeroes in it.

The problem is that the IT industry has been painfully slow to get behind IPv6. That's understandable, since legacy networking equipment rarely supports the standard and often lacks the ability to add support via firmware upgrades. Even today, many Internet hosts and routers have serious problems with IPv6 compatibility.

So migrating to IPv6 is an expensive, time-consuming solution to a problem that always seemed just a little too far away to take seriously.

That distance, however, turned out to be an illusion: We're very close to hitting the wall with IPv4. Lots of Web sites now post "IP exhaustion countdown" widgets that purport to show how much time remains until the big day.

But it won't really be a "big day." Since IP addresses are allocated in a hierarchical manner, IPv4 exhaustion will happen in a series of smaller steps. Eventually, however, ordinary business Internet users are likely to feel a financial pinch as the IP address shortage trickles down.

Geoff Huston is chief scientist at APNIC, one of the regional Internet registries that distributes IP addresses to ISPs and other large organizational users. APINC and other regional registries get their IP addresses from the Internet Assigned Numbers Authority (IANA). According to Huston, IANA will run out of addresses some time between July 2009 and 2011.

APINC and other regional registries will, in turn, exhaust their supply of IPv4 addresses between October 2009 and July 2012.

What happens next? At first, perhaps nothing. Many ISPs and large organizations are sitting on large blocks of unused IPv4 addresses. Unless the rules change considerably (and they just might), these groups will be free to redistribute the unused addresses as they see fit.

This brings us to the next phase of the address exhaustion process: the point where ordinary business users that rely on IP addresses to host Web sites or to run servers may notice some changes.

David Siegel, a VP at Global Crossing, writes that "This could be a very interesting time, as ISP's that have space to allocate may have a competitive advantage over those that do not and may be able to charge a premium as a result!"

In other words, service providers will have every incentive to charge more -- perhaps much more -- to provide users with static IP addresses. According to Siegel, we could see this start to happen within 4-6 years.

Finally, Siegel writes, we will reach the final step in the IPv4 exhaustion process: "The last and final phase, much like the selling of datacenter space, will involve money and perhaps some begging. The black market, perhaps Ebay, will become the only place that one can acquire addresses. They may not even be transferable across upstream providers, but that won't keep people from trying to trade them. Eventually the cost of IPv4 address space will be more than someone is willing to pay for and the v4 Internet will officially be sold out." Seigel thinks this phase will take up to 10 years to play out. Of course, companies that rely upon static IP addresses will be paying through their noses quite a bit sooner than that.

Will the IT industry ride to the rescue and push to get enough IPv6 infrastructure into place before it comes to this? A so-called "dual stack" strategy exists to migrate the world from IPv4 to v6, and it could allow a smooth, trouble-free migration path.

Or at least it could have if we had started the process about a decade sooner. Here is what Huston writes about the dual stack strategy: "Under this dual-stack regime every new host on the Internet is envisaged to need access to both IPv6 and IPv4 addresses in order to converse with any other host using IPv6 or IPv4. Of course this approach works as long as there is a continuing supply of IPv4 addresses, implying that the envisioned timing of the transition was meant to have been completed by the time that IPv4 address exhaustion happens." The catch: Huston concludes that even if we had started working seriously on a migration process in 2008, we might still need twice as many IPv4 addresses to complete it without hitting any IP address shortages.

Now, it's a moot point. The industry waited too long to get moving, and it will pay the price. The only questions are how much and for how long.

Every IT problem tends to create a new market for solutions. This problem is no different: So-called "tunnel brokers," for example, provide a certain measure of interoperability between IPv4 and IPv6 networks.

These technologies, however, can introduce new problems of their own. Also, inevitably, such solutions represent a new source of cost and complexity for businesses that implement them.

This is a fascinating problem, and it illustrates how big, slow-moving, supposedly risk-averse organizations still get clobbered by problems they see coming miles -- or years -- away. But will there be anything ordinary businesses can do about it, except watch this mess unfold and hope it doesn't cost them too much?

I doubt it. IP exhaustion isn't the end of the world by any means. But it could make the Internet a more expensive place to do business for a long time to come.