For Cisco, It's Change Or Perish - By Bret Swanson - 04.18.01

For Cisco, It’s Change or Perish By BRET SWANSON April 18, 2001; Page A22

Oh, how the mighty fall. It wasn't so long ago that Cisco Systems was the world's largest company and the envy of all the technology industry. It had exceeded published earnings estimates 14 straight quarters in a row, was hiring like crazy, and had topped $555 billion in market capitalization. These days, Cisco is slashing 15% of its workforce, missing all its numbers and wondering where more than $400 billion of market value went. To his credit, Cisco head John Chambers saw things coming and was the first chief executive to smartly call on the Federal Reserve to cut interest rates and for Congress to cut tax rates. But unfortunately for Cisco, Washington getting its act together won't be enough to put the networking company on the rebound. That's because Mr. Chambers's real problems aren't economic. Cisco is increasingly in the wrong business. Heroic Processing The company that defined the 1990s and made the Internet possible is being squeezed from above by fiber optics and from below by microchips. It has a core competency in neither. Cisco's great advance was its vision for the first post-telephone network. It developed the architecture and crucial software standards that made the Internet work. Then it cobbled together the chips and wires and code into high-margin boxes that few others knew how to make. But unless Mr. Chambers can set his firm on a new path, Cisco will be remembered as a transitional, if temporarily transformative, enterprise. The key to understanding Cisco's prospects is to understand the history of communications networks. Real-time communication over a distance was first made possible by Samuel Morse's telegraph, or copper wire carrying electromagnetic signals. As technology evolved we were soon able to send voices, rather than just simple dashes and dots. Voice calls were switched by operators plugging wires into a switchboard. Later, this switching function would be overtaken by mammoth electronic boxes from Lucent and Nortel.

Still, the architecture remained the same for nearly a century: Each telegraph or phone call received its own connection. You may have only used your telephone 20 minutes a day, and you may have only spoken 15% of the time, but the connection was always there. Relative to the application, we had a lot of bandwidth to spare. Enter the Internet. With the first online services in the early 1990s and the introduction of the Netscape browser in 1994, the data-centric Internet became a phenomenon. E-mails, Web browsing, and later e-commerce threatened to overwhelm the voiceoptimized telephone network. The mostly copper telecommunications infrastructure wasn't ready.

Broadcom, Applied Micro Circuits, EZchip and others are integrating more and more of the functionality of Cisco boxes Cisco foresaw this data explosion. And it found a way around the narrowband copper links to offices and homes, as well as the more capacious -- but still not big enough -- fiber-optic strands connecting our cities. Cisco discovered that if you didn't give everyone a connection, but instead chopped up the messages and gave the packets a destination, you could throw the packets out into the Internet, let them take any number of paths, and then have them reassemble at the other end. The new Cisco-powered network was ruthlessly efficient. There were no more circuits between users. There was no more silence. The Cisco network used every bit of bandwidth available. Now, bandwidth scarcity is over and Cisco's time has passed. The new all-optical network doesn't need heroic processing of billions of bits each second. All it must do is change the direction that light travels. The Cisco network treats messages as bits rather than lightwaves, laboriously processing what should simply be steered. It's as if rivers worked not by banking the flow of the stream but by checking and sorting each H2O molecule for its KMI reports that in 1999 and 2000 some 150 million kilometers of optical fiber were laid world-wide. Avanex, the leading optical networking

company, can make and sort and direct thousands of colors of light, each carrying distinct messages, down a single fiber strand. Yipes is connecting urban business directly to fiber, and wireless start-up Soma Networks can supply residential broadband connections that are faster than copper DSL or broadband cable TV. Cisco and its big, central routers can't handle this second wave of data. And right now it has no back-up strategy. Focused for too long on routers and other electronic network modules, Cisco doesn't now have the optical capabilities to prosper in a new world of bandwidth abundance. Just weeks ago Cisco announced it would shut down one of its lone optical plays, Monterey Networks. Commenting last May, we noted in the Gilder Technology Report that there were too many real optical companies for Cisco to simply dabble: "Cisco's $500 million investment in Monterey is suddenly irrelevant and will never be a significant factor in the network. . . It's Sonet hardware investment, $7 billion for Cerent, may be even more fruitful than anticipated in the short term . . . but faces an even shorter life expectancy." Indeed, while Cerent was a revenue star in 2000, in 2001 it has already been rendered obsolete by the continually multiplying colors of light from Avanex, ONI, and Sorrento. Creative Destruction Of course we will still need electronic terminals and routers and switches at the edges of the network. But if Cisco's failing optical strategy weren't enough, its products are increasingly being disrupted by Moore's Law of microchips. Broadcom, Applied Micro Circuits, EZchip and others are integrating more and more of the functionality of Cisco boxes -- all the wires and separate processors and software -- onto single silicon chips. If the router becomes a chip, what does Cisco have to sell except its software and protocols? The destruction in the market affords Mr. Chambers a perfect opportunity to practice some creative destruction on his own company. Namely, he needs to remake it in the image of the new all-optical network. Mr. Swanson is a technology analyst at the Gilder Technology Report.