TELECOM: CAN THE OLD GUARD SWITCH GEAR?
A telephone-company switch looks almost like a mainframe computer. It's a huge metal cabinet in a locked room, costs millions of dollars, and has thousands of wires leading out of it. The companies that make them--such as AT&T, Northern Telecom, Siemens, Alcatel, and Fujitsu--have noticed the similarity. And given the mainframe's sorry state, the comparison terrifies them. Says Harry Newton, a New York telecommunications consultant and publisher: "This is the last and largest major industry that's reassessing every assumption."
The traditional manufacturers of telecom equipment are about to get zapped by the new rules of the game. They're accustomed to product cycles lasting 20 years. They are used to making big profits from software upgrades that no one else can supply. And the only things they give away are canapes and cocktails to phone-company purchasing managers.
To be sure, the switching business is still healthy. AT&T's Network Systems unit posted a 20% gain in revenue last year. China alone anticipates buying switches to serve 10 million new phone lines a year. And change tends to come slowly: Phone service is too vital to be trusted to the latest untested "hot box."
WAVE UPON WAVE. For all that, old-line companies smell change. They know that their products are vulnerable to the same punishing cost curves as computers. Plentiful capacity over optical fibers makes new types of equipment feasible, while deregulation is luring in new service providers that are willing to try different suppliers. "You have three or four waves of technology coming, all in the context of multibillion-dollar investment plays," says Patrick Courtin, CEO of ComStream Inc., a San Diego-based manufacturer of digital transmission gear.
AT&T and its ilk have already largely missed the booming market for data-networking gear from companies like Cisco Systems, Cabletron Systems, and Bay Networks. The old guard thinks it can recoup lost ground with asynchronous transfer mode (ATM), which is capable of mixing voice, data, and video traffic. The big switchmakers have poured hundreds of millions of dollars into ATM development. Yet so far, most sales have been in office networks, where smaller companies, such as Newbridge Networks Inc., have gobbled up the orders. Finland leads the world in per-capita deployment of ATM in public networks, yet it has bought virtually none of its gear from the big-name players. Instead, Juha Heinanen, technology director for Telecom Finland, has installed smaller ATM switches from such companies as General DataComm in Middlebury, Conn. Says Heinanen: "You buy a big switch, and it's obsolete almost instantly. A small one, you can find a place to move it where the new requirement isn't so important."
"DO YOU CARE?" Lightly populated Finland may not be a typical case. Yet the core business of circuit switching is also under attack. Dozens of small companies are building cheap switches from the innards of personal computers. They're slightly less reliable, but that may be O.K. for certain price-sensitive applications. Meanwhile, Microsoft, IBM, and Novell are pushing standards that give computers control over costly office phone switches called PBXes. Once the computer is in charge, asks Russell W. Bell, a marketing director at Advanced Micro Devices Inc., "do you care if the PBX is made by AT&T or XYZ?"
To be sure, there will always be a role for large, complex switches of the type made by AT&T, Siemens, and the like. Plus, those giants retain unparalleled expertise in managing network design and construction. They also understand how to guard network reliability. Says Harry Bosco, the top engineer for AT&T Network Systems: "You can't just throw it all in the air and say, `We'll change the whole direction overnight."' True enough. But cheap and fast always seem to win out. If these giants don't move quickly, they may end up following someone else.By Peter Coy and Neil Gross in New York