Advertisement
YOU ARE HERE: LAT HomeCollections

TRENDS | INNOVATION / STEVE G. STEINBERG

Addressing the Internet's Space Problem

August 26, 1996|STEVE G. STEINBERG | Steve G. Steinberg (steve@wired.com) is an editor at Wired magazine

The Internet is running out of space. Every computer connected to the Internet must have a unique name and address, but there are only so many available.

On the name side, this has already become an inconvenience. Salon magazine, for example, ended up with salon1999.com because salon.com had already been taken by a hair salon in Austin, Texas. Some speculators have even taken to registering domain names by the thousands on the theory that eventually somebody will want most of them and be willing to pay for them.

But it's the looming shortage of addresses that spells real trouble.

An Internet address is a unique string of 32 1s and 0s. That's enough bits to identify many billions of computers. But back when the Net was young and billions seemed like infinity, addresses were given out in huge blocks by the Internet Assigned Numbers Authority, or IANA. Rather than give, say, MIT one address at a time, IANA would give it a block of 16 million addresses to use as the school wished.

That sort of thing added up. Today, only a small percentage of Internet addresses is actually in use, but most of them are spoken for. Now IANA must be positively miserly with the address blocks it hands out. Its people look closely at every organization that applies for a block and try to predict how many addresses will really be needed.

The results are highly subjective, and sometimes controversial. For example, TCI's Internet cable modem service @Home was recently assigned a much larger block of addresses than normal. That stirred up envy among smaller Internet service providers. After all, a huge block is a competitive advantage for @Home because it will allow it to better support a large number of customers.

The obvious solution to this address famine is to increase the length of the addresses. Why not have 128-bit addresses instead of 32-bit? In fact, a new version of the Internet Protocol (IP version 6) has been developed that does just that. But a couple of problems have stalled its adoption.

First off, switching over to IPv6 will be incredibly laborious. It's roughly equivalent to making every phone number in the United States 14 digits long. Every phone book would be instantly out of date, and the entire phone system would have to be reprogrammed.

Second, the Internet's routers--special computers that direct data to the correct destinations--are having a tough time keeping track of all the addresses that exist today. Increase the number of addresses, some fear, and the Internet's routers will be brought to their knees.

So if Internet addresses are doomed to be a scarce resource for at least the next few years, what can we do to make sure they are used efficiently? The answer, according to a growing number of voices, is to bring market forces to bear.

Two months ago at an Internet conference in Montreal, a working group was formed with the goal of creating a market for buying and selling Internet addresses. Whereas today Internet addresses are free and can only be obtained from IANA or one of three regional authorities, an address market would allow all who have address space to sell portions of it directly.

Such a market, says Paul Resnick of AT&T Laboratories, would provide financial incentives for organizations to sell off addresses they own but don't really need. At the same time, companies would no longer try to obtain larger address blocks than absolutely necessary, because they'd have to pay more.

And a market would eliminate the perceived unfairness of the current scheme. Instead of some central authority deciding how much an organization deserves, it would simply be a question of how much the organization is willing to spend.

Creating a market is a pretty simple idea, and we know it works. It's how other scarce resources are managed, from real estate to gold. But for the Internet community, which has long relied on goodwill and government funding to keep everything running, it represents a radical shift in philosophy. Turning to financial incentives so that people will do the right thing is an admission that the old days of selfless cooperation are over.

Some critics even see the shift as evidence of a power grab. According to Bill Manning, a researcher at USC's Information Sciences Institute, the effort to create an Internet address market is being driven by AT&T. A recent request by the company for additional address space was turned down. Now, he says, AT&T is trying to wrest control from the assigning authorities so it can buy the addresses outright.

This seems a touch paranoid--plenty of other people are involved in the effort--but there is no doubt that big players like AT&T stand to gain the most. They will be able to buy what they need no matter how costly. But as AT&T's Resnick points out, "People with money always have advantages in cases of scarcity." That's just a fact of life, albeit one that the Internet has been able to isolate itself from up until now.

Besides, argues Resnick, the scarcity of addresses is really only a technological artifact, so we don't have to worry about address hoarding or price manipulation. If the cost of Internet addresses becomes out of reach for small companies or schools, it will act as an incentive for people to finally move to IPv6.

There's also a wild-card factor that could ultimately prevent a giant corporation like AT&T from seizing control of the Internet's addresses: Nobody is sure who owns them. The Defense Department has made rumblings in the past saying it owns the entire address space because the space is the product of research it supported. But nobody is sure if that claim would really hold up, and nobody is particularly eager to find out. As long as it's unclear who has control, it makes it very hard for anyone to grab it.

Advertisement
Los Angeles Times Articles
|
|
|