The Internet Puzzle

Who will hold the biggest pieces?

Dont give up on the Internet or the carriers jockeying for a piece of its traffic just yet.

The Internet is growing faster now than it did in the halcyon days of 1998 and 1999, and the race to decide which company will carry the traffic is much closer and more crowded than most believe.

Internet traffic quadrupled last year and is still keeping that pace, which is faster than the stunning 280 percent average annual traffic increase recorded in the late 1990s, says Lawrence Roberts, who 30 years ago created ARPANet, an ancestor of the Internet.

Roberts got unprecedented access to information about the data-carrying capacity of 19 top carriers, but with the caveat that he not discuss the carriers by name. He found that the company carrying the most traffic has a 14 percent share — and three others are just behind it with about 12 percent each. The next four have shares of 5 percent to 8 percent, and five others have more than 3 percent of the pie.

"Its a much flatter distribution than I expected," says Roberts, now chief technology officer of Caspian Networks, an optical switch vendor. "Its very different than most people assumed. You talk to carriers and they think theyre way below the top players, because they think the top carrier has 40 [percent] or 50 percent of the traffic."

With the race so close, carriers and service providers "had better start buying equipment soon to maintain their market share," he says.

Roberts predicts that Internet traffic will continue to quadruple for several years. Carriers able to maintain their market share should be able to take advantage of steep drops in the cost of bandwidth and gear, and double their revenue yearly, he says.

Yankee Group analyst Nancee Ruzicka generally agrees with Roberts findings, and says her conversations with carriers substantiate his forecasts. "Spending will pick up," she says, but she cautioned that because carriers are pushing for such steep discounts from vendors, the revenue for equipment vendors might show only a modest uptick.

The Yankee Group expects the core router market to grow 64 percent each year through 2003, to $11.1 billion from $3.7 billion this year.

Fewer Players

Certainly, there will be some consolidations and failures — and there wont be as many players in 2003 as there are now, Roberts says. Still, "Its wide open now. Its basically a horse race between eight or 12."

Roberts expects revenue from carrying data across the network will exceed voice revenue in two or three years. He compared carrying capacity on trunk ports for April 2000, October 2000 and April 2001. The estimates were based on the traffic during average and peak days. The average load on an average day was 40 percent of capacity. On a peak day, theyre likely to be at 95 percent capacity, and at 99 percent during a peak minute. So, if a carriers average day is at 40 percent capacity and it acquires new customers, more ports are needed.

Some time in 2000, the corporate Internet market in the U.S. hit critical mass, with businesses realizing that they had to make it easier for customers to buy, trade and exchange information through Web sites or theyd be left behind, Roberts says. That accelerated traffic here. Europe and Asia havent reached that critical mass yet, but when they do, there should be another explosion as firms overseas scramble to boost their bandwidth.

"Companies havent reached the point where they have to use the Internet all day, every day," he says. High-speed bandwidth delivery could top out at about $100 billion per year in nine more years, he adds.

Some competitive carriers, such as Broadwing, Time Warner Telecom and XO Communications, have the customer base to hold their own against the biggest guns, such as Qwest Communications International and WorldComs UUnet. Others have filed for bankruptcy. Competitive carriers tried to build nationwide networks as fast as they could, but most never got the customer penetration rates necessary to sustain the mammoth capital outlay.

How has traffic kept skyrocketing despite the crashing and burning of so many dot-coms, stocks and competitive carriers? Wall Street overreacted to the collapse of the business-to-consumer market — which never accounted for more than 5 percent of Internet traffic, Roberts says.

Intel CEO Craig R. Barrett agrees, noting that on a typical day his company sells $60 million worth of products over the Net to other companies. "The Internet will be the driving force of the computer and telecommunications industries. Business-to-business is what is going to drive Internet commerce," Barrett says.