In August 1999, Cisco Systems completed two acquisitions to create a new business in the optical networking arena. Cisco purchased Monterey Networks for 7.3 million shares ($500 million) of Cisco stock, and Cerent for 100 million shares ($6.9 billion). Analysts expect this move to strengthen Cisco's position to expand its product line into broadband technologies such as high-speed access for home users, wireless networking, and security.
Monterey's optical cross-connect products and Wavelength Routers let ISPs concentrate signals to increase network capacity. Cerent makes switches that let telephone companies route, concentrate, and accelerate Internet data, then feed that data to fiber-optic lines designed for voice traffic. Monterey's optical products and Cerent's switches will let Cisco expand its business from LAN routers and switches to optical network switching. Carl Russo, Cerent's CEO, said, "This technology gives Cisco direct access to a marketplace that is huge." Cerent's executives project that by 2002 the overall switching market might reach $17 billion and the optical transport market might reach $10 billion.
Cisco is buying its way into a leadership position in optical networking, partially as a hedge against the increased competition with its main line of business (LOB), and because Cisco recognizes that high-speed data transfer over fiber will become a major business. Intel has already taken direct aim at Cisco with its new networking-chips business. The optical networking market is no less competitive: Lucent Technologies and Nortel Networks have significant positions, and Siemens and Alcatel are also active players.
Cerent's price was steep for a company of Cisco's size. Because both of the companies that Cisco purchased had small revenues the previous year, John Chambers, Cisco president and CEO, had to defend the purchases. At the first Dell DirectConnect conference in August 1999, Chambers said, "Anyone who understands our market understands that the optical transport business is going to explode. The companies that can't partner and acquire will be left behind. Getting everything connected is everything." Chambers noted that he expects the wireless network technology to similarly explode during the next few years.
Analysts predict that by 2010, half of the US economy's transactions will flow over the Internet. Beyond serving transactions, optical networking will require video, audio, and expertise. High-capacity pipes will handle massive volumes of voice and data together. The amount of transoceanic cable laid worldwide has dramatically increased during the past few years. A small group of analysts believe that broadband infrastructure will be overbuilt, but foreseeing how that could happen is difficult because technology seems to expand to fill any size pipe. The backbone of broadband technology will be optical wiring, and fiber will eventually run into individual homes and businesses. You might think that satellites will compete with cable for backbone technology, but satellites haven't competed thus far.
In Europe, many innovators already have optical projects under way through Cambridge University. If the future of networking is optical, as these innovators believe, let's pause to consider the implications of an outage of a main artery of the broadband network. The United States has already seen the havoc that a cut into a main fiber backbone of New York City caused a few years ago. Add Internet traffic to the outage, mix in companies that depend on Internet sites to provide 24 * 7 uptime, and you have the frightening prospect of bringing vital enterprises in diverse geographical areas to their knees. If this scenario is correct and network connectivity convergence reduces redundancy and provides larger points of failure, you can expect to see more fiascos like the outage in New York in the near future.
The rise of optical networks is likely to change everything we know about TCP/IP networks. In the July issue of The Gilder Technology Report, George Gilder argues that optical networks will relegate IP to a minor technology—if not kill IP. Gilder also notes that progress in optical transmission and switching technology will lead to progress in fiber-wave-division multiplexing.
Gilder projects that fibers will eventually transmit 10Gbps on each of 500 wavelengths (lambdas) of 140 fibers in a cable. This transmission capacity would mean a throughput of petabits (1015 bits). From this scenario, Gilder forecasts, "Listening to the technology, one can hear that the all-optical network won't switch packets, it'll shuffle wavelengths."
The shuffling of wavelengths means a return to circuit-switching technology based on lambdas. So, Cisco's purchase of the technology that transforms and concentrates electrical signals from standard Ethernet cables into light might have great strategic value.
The breakthrough in fiber-optic performance and its implications for broadband networking might be one of the greatest achievements of the late 1990s. Bob Herbold, Microsoft executive vice president and CEO, said, "In 1996, fiber's capacity was 10Gbps. In 1998, the capacity was 40Gbps. In the labs, fiber is transmitting 3000Gbps. Therefore, a next-generation network could transmit all the telephone calls in the world \[i.e., approaching 17,000Gbps\] over six strands of fiber."
No wonder Cisco is willing to pay billions now to get in on the ground floor of this new technology. If TCP/IP infrastructures are obsolete in 5 to 10 years, as Gilder suggests, the high price Cisco paid for its entry ticket will pay off handsomely.