Full Speed Ahead on a High-Speed Internet

At the height of Web mania, financiers and engineers agreed that a faster Internet would be a building block of growth — and a great investment as well. So does Web mania spell the end of that high-speed vision? Not so fast.


If there ever is an Internet Hall of Fame, you can bet that Pradeep Sindhu will be elected on the first ballot. He is a former Xerox engineer who in 1995 came up with a radically better way to handle high-volume traffic along the Internet backbone. That vision became Juniper Networks Inc. — a startup that quickly grabbed more than one-third of the high-end router market, becoming one of the few Internet-equipment companies ever to compete successfully against industry leader Cisco Systems.


The past year has been brutally tough on Juniper. Some of its best customers are stretching out orders or struggling to stay solvent. Juniper’s stock, which last year topped $200 a share, skidded this autumn to less than $15 — and all that before the terrorist attacks on the United States.

So is Sindhu discouraged?

No way. In a recent interview, he argued that the case for a high-speed Internet is stronger than ever. “Ideas are flowing much faster now than they ever used to,” he observes. “There’s enormous value in having any computer capable of connecting to any other computer. We’re building the cerebral cortex of the real world. And that will help us solve problems that we haven’t even formulated yet.”

Just look at Juniper’s own engineering teams to see this potential in action, he adds. “We’re very tight with IBM,” he explains. “We design chips on the West Coast, and IBM manufactures them back East. We send designs — hundreds of megabits at a time — back and forth constantly. Two or three years ago, it was infeasible to do that. Now information flies back and forth as if barriers to distance and time have been erased.”

Think more broadly, Sindhu says, and it’s easy to envision vast areas of our lives that a faster Internet can improve. Interactive video could be crisp, quick, and effortless. Complex equipment, including airplanes, could be controlled via the Internet. “We’re just at the beginning of our applications,” he says.

As Juniper’s vice chairman and chief technology officer, Sindhu is mindful of the technology sector’s current slump, and he’s watching European and Asian markets to see how much the U.S. economy’s troubles may spread. Nonetheless, in his view, such short-term jitters are just that — short term.


Once the current hard times are over, he says, Internet traffic and bandwidth should resume breakneck growth. In the past few years, growth rates have averaged 100% a year or more. It’s anyone’s guess how much that rate will drop for full year 2001 and the start of 2002. Looking further out, Sindhu contends, 80% to 90% annual growth could continue for some time to come.

Think a bit about Sindhu’s perspective — and all of a sudden there’s reason to rethink one of the most morose bits of current conventional wisdom. Yes, this has been a dismal year for almost anything connected with the digital economy. With bad news all over the headlines, it’s easy to ask, “Can we really afford to build a faster Internet?”

But maybe the right question is just the opposite: Can we afford not to build a faster Internet? Consider some of the usage forecasts made recently by research firms, even after it had become clear that U.S. demand this year had moderated.

By 2005, according to International Data Group of Framingham, Massachusetts, there will be at least 240 million Web users in the Asia/Pacific market. That would surpass the U.S. total, with China, India, and Korea leading the pace. And even after that boom, barely one-tenth of Asia will be online.

In the United States, high-speed access is as alluring as ever — even though the DSL and cable industries haven’t made it easy for people to get connected. The Yankee Group recently predicted that some 15.7 million households will have broadband cable Internet access by the end of 2005, while 10.5 million will have DSL, 4.5 million will have satellite access, and 359,000 will opt for wireless access. All of those numbers represent at least a doubling of current levels.

What will people want to do with those high-speed connections? The film industry is betting that downloadable movies-on-demand will take off. Major studios have already formed two consortia — Moviefly and — to develop such services. Napster-style downloadable music ran into all sorts of legal snags this year, but it’s a good bet that eventually the recording labels, artists, and consumers will settle on some form of payment that is tolerable to everyone.


Perhaps most important, business appetite for bandwidth is likely to keep growing for the foreseeable future. Email attachments are likely to keep getting bigger — especially in creative and technical fields. Online video is likely to become an ever-more powerful business tool, instead of just a source of entertainment. Within a few years, high-resolution video over the Internet could become an essential, everyday part of e-learning — whether it’s for university credit or just to keep your sales force up to date.

As new services are developed, business connection speeds that seem fast today may seem laughably slow in a few years. The same goes for Internet architectures. As a matter of fact, Juniper itself is combining its traditional obsession about speed with an expanded, new focus on building more intelligence and smarter routing into the network as well.

One of the most intriguing perspectives on Internet speed comes from Michael Ritter, the former CTO of San Jose-based Metricom Inc. His company developed the Ricochet wireless modem, which provided Internet connections at 128 KBPS or greater to about 51,000 users throughout the United States. Devotees loved it — but the company stretched itself too thin in its effort to build a nationwide network and was forced into bankruptcy this past summer.

“We had a huge retail following,” Ritter recalls. “Engineers who installed the service liked it so much that they often ended up buying our stock. Near the end, we figured out how to double our speed, to 256 KBPS, just by changing the software. And we had ideas in the lab that would have made it 10 times as fast. We just misread the signals from Wall Street. We spent almost all of our money on a national expansion plan when we should have stayed in a few markets and grown more slowly.”

Plenty of other innovators are grumbling about the fickleness of the financial markets. For a while, Wall Street’s best thinkers encouraged even the most audacious, pell-mell spending plans — and then investors abruptly stopped tolerating big losses in the name of growth.

But there’s an old-time Wall Street saying that captures the current situation with eerie precision. “You’re never wrong,” the stock-market adage goes. “You’re just early.”


George Anders ( , a Fast Company senior editor, chronicles the present and the future of the Internet from his base in Silicon Valley.