PREMIUM SEARCH Search by job title, geography and build a list of executive contacts
Some entrepreneurs have their hopes pinned on a roaring IPO or the chance to build a billion-dollar business from scratch. Others have already tasted success or failure and dream of starting all over again. Then there's Judy Estrin -- perhaps the most prolific serial entrepreneur in computer networking and one of the most visible, successful female executives in the technology sector.
Along with her husband, Bill Carrico, Estrin has launched three networking companies over the past two decades that grew to be worth hundreds of millions of dollars. On June 12, the couple kicked off their fourth startup, dubbed Packet Design Inc., with a novel business model and an audacious agenda. The goal is nothing less than saving the Internet from technological paralysis. "There's a vacuum out there right now," says Estrin, who in April stepped down as chief technical officer for networking giant Cisco Systems Inc. after two years on the job. "We're filling the research gap between industry and academia."
The concern Estrin perceives is little understood outside technical circles, but it has the potential to throttle the Net's future growth. The problem: Companies such as Cisco are so focused on near-term product cycles and technical patches that grand engineering challenges, such as how to redesign the Net to carry video or to handle billions of new users, aren't getting solved. Meanwhile, research institutions face declining funding and are often too far removed from practical networking concerns.
BRIGHT LIGHTS.
To bridge the gap, Estrin and Carrico have lined up an all-star cast of staff, investors, and advisers in a new kind of business incubator. Rather than creating and selling its own branded technologies, Packet Design, based in Menlo Park, Calif., will solve specific networking software problems and then set up separate businesses to sell the solutions.
Specifically, Packet Design will dream up improvements to Internet protocol, the Net's lingua franca, and then hand them to spin-offs in which it retains equity ownership. The first should happen within two years. "We're not a product company, but a technology company," says Estrin, 45, president and CEO. Carrico, 50, is chairman.
The whole idea might seem like just a dream if not for the luminaries who have signed on to help. Packet Design has raised $24 million from venture firm Foundation Capital, as well as from Estrin, Carrico, and individuals such as Bill Joy, a co-founder of Sun Microsystems, and James Barksdale, former CEO of Netscape Communications. Its advisers include Vinton Cerf, one of the fathers of the Net and now a top exec at MCI WorldCom Inc., and Harvard University networking guru Scott Bradner.
MIRED IN POLITICS.
But perhaps Packet Design's biggest recruitment coup is chief scientist Van Jacobson, former chief scientist at Cisco and prior to that a networking honcho at Lawrence Berkeley Laboratories. Jacobson's huge contributions to Net technology are nearly equaled by his anonymity. Among other things, he has researched technologies for "multicasting" video across the Net and for controlling Net congestion.
Now he and the rest of the Packet Design crew face a different set of hurdles. Currently, under the auspices of the nonprofit Internet Engineering Task Force (IETF), a vast volunteer collective devises new Internet standards. But as the Net has exploded, the IETF has become increasingly mired in vendor politics and unable to effect big changes. "It's a body aimed at resolving ideas brought to it by companies," Carrico says. But, he charges, that can't happen if members don't do basic research on problems such as how the Net will accommodate millions of handheld wireless gizmos. "Most of the proposals before the IETF today are short-term solutions."
Packet Design doesn't aim to supplant the IETF -- that would be suicidal in the Net's collaborative environment -- but to speed the rate at which new ideas reach its committees. Or, if necessary, to bring breakthrough technologies quickly to market and let standards bodies race to catch up.
"LIFESTYLE ISSUE."
Such out-of-the-box thinking is old hat for Estrin and Carrico. Their first company, Bridge Communications, was an early pioneer in routers, the specialized computers that direct traffic around and between networks. Founded in 1981, three years before Cisco, Bridge merged with 3Com Corp. in 1987 in a $235 million deal.
Estrin and Carrico went on to found Network Computing Devices, a maker of high-resolution computer terminals for UNIX systems in 1988. Then, in 1995, they started Precept Software, which made multimedia networking technology. Precept was acquired by Cisco for $84 million in 1998, and Estrin became the company's CTO. Carrico came on board to run Cisco's small- and medium-size business market group.
But both eventually itched to get back to a startup. "I really like building companies," Estrin says. "It's a personal-lifestyle issue." With its esoteric mission and unusual structure, this latest Estrin-Carrico production may not be a world-beater in name-brand technology or eye-popping value creation. But it's aiming at a higher purpose -- one Estrin and Carrico can legitimately aspire to achieve at this stage in their careers.
Packet Design intends to save the Net from choking on its own success. After helping drive the whole phenomenon in the first place, networking's First Couple now are turning their backs on the tyranny of warp-speed "Internet time" and the search for a quick payback. If it works, the results could make them heroes of the 21st century Internet.
Reinhardt covers technology for Business Week from San Mateo, Calif.
EDITED BY BETH BELTON
Get BusinessWeek directly on your desktop with our RSS feeds.
Add BusinessWeek news to your Web site with our headline feed.
Click to buy an e-print or reprint of a BusinessWeek or BusinessWeek Online story or video.
To subscribe online to BusinessWeek magazine, please click here.
Copyright 2000 , by The McGraw-Hill Companies Inc. All rights reserved.
Printer-Friendly Version
