(FORTUNE Magazine) – As his whirlybird settles on the landing pad beside Lucent's Darth Vader-esque New Jersey headquarters on a golden, still April afternoon, CEO Richard McGinn, 51, is surely the master of his domain. Since he took over last October, revenues have surged and the stock price has nearly doubled. Just this day he has unveiled a $1 billion cash purchase of a small technology company that should help Lucent shed whatever vestiges of Bell-era sluggishness that might cling to it, and play in the brash new world of the Internet.

Buoyed by an unexpected resiliency in demand for its old-world telecom gear, Lucent Technologies (1997 revenues: $26.4 billion) has been lucky since its 1996 spinoff from AT&T. Its biggest customers, the Baby Bells, continue to crave telephone switches and transmission gear as demand for Internet and fax connections fuels an explosion in telephone lines. New local phone companies are also buying the stuff, as are phone companies installing voice service for the first time overseas, especially in China. The growth in the conventional voice-based business has been a pleasant surprise, says McGinn: "Every quarter for the last seven quarters I have been unwilling to say that it is a trend." Lucent has also done a great job selling the infrastructure for cellular phone systems, in one case stealing a huge contract from industry leader Motorola.

The old world won't last forever, though. As they struggle to meet the demand for new lines, phone companies are frantic to shunt data traffic from their overstretched networks, which were designed to carry voice, not the ones and zeros of computer language. That means Lucent must be ready to move with them into the new realm of data networking. As Lucent talks to big customers about the network of the future, it is butting up against a new enemy everywhere it turns: Cisco Systems. Says Lucent networking chief William O'Shea: "We see them in almost every major bid."

Cisco is a much smaller company; in fiscal 1997 it sold a comparatively measly $6.4 billion of gear, mostly routers that move data across the Internet, through corporate networks, and between companies exchanging information. But its 1997 sales were up 57%, vs. a mere 13% increase for Lucent. If Lucent is the top dog in telephone network equipment in the U.S., handily beating its nearest rival, Northern Telecom, Cisco is the Godzilla of datacom, towering over rivals like Bay Networks and 3Com.

Lucent and Cisco are locked in a contest to define a new class of network based on the rules of the Internet. Today heavy users like corporations route most of their data traffic--not just faxes but valuable stuff like blueprints and legal documents--on expensive private lines leased from phone companies. The Internet, and other new networks based on its rules, could save them money by letting that data commingle with traffic from many other users. But for corporations to entrust them with their lifeblood, data networks have to be invested with the same sort of reliability taken for granted with telephone networks.

These reinvented data networks eventually will carry voice and video communications too. Today the volume of data traffic on the world's communication networks is roughly equivalent to the volume of ordinary calls. But data traffic is growing wildly, somewhere between 200% and 600% a year, vs. just 10% a year for voice. Within five years or so, the volume of traffic on data networks will dwarf that of phone calls. Then it will make sense to start converting phone calls into digital bits and adding them to the traffic flow at very little cost, like fleas hitching a ride on an elephant.

The incentive is enormous for phone companies to get a piece of that action, especially since people using the public phone network for data--by sending faxes, say, or using PCs with modems--tie up costly telephone switches much longer than they would a data network. Besides, running data networks would let phone companies get into lucrative new fields like E-commerce and electronic publishing, which offer the potential for greater profits than, say, selling another minute of telephone time for a nickel on Mother's Day.

The trick for telephone companies is to build the network of the future while keeping up with traffic growth on the network of the present. Lucent wants to provide the nuts and bolts for this transition. But the company does not yet make great products for moving data over networks based on the rules of the Internet. Cisco, on the other hand, has made its fortune building private data networks for corporate customers. Although that sector is still strong, growth has slowed recently while competition has increased, so Cisco is turning its attention to the customers demanding more equipment than ever--phone companies. That's why it has lined up for this vicious scrum against Lucent.

Neither company has the product it needs to ensure a big win. Cisco, too, needs to modify its products to suit the new world. Routers are the data-network equivalent of a telephone switch, identifying the destination of an incoming packet of data and sending it off in the right direction. But routers don't provide the on-time guarantees that phone companies demand. Data like E-mail messages wend their way to their destinations on trips of uncertain duration. That sort of imprecision won't do for phone calls or video signals. So all the networking companies--including Lucent, Northern Telecom, 3Com, and Bay--are racing to deliver a hybrid device with the speed and efficiency of a router and the precision of a telephone switch. The class of equipment is so new that Lucent executives sometimes aren't sure what to call the device they are about to introduce--a switched router or a routing switch. Whatever its name, the thing will identify high-priority traffic and deliver it immediately.

Lucent and Cisco each have different talents they will use to woo the phone companies. Lucent starts with a century of experience making equipment for the telcos, and its biggest selling point with them is the rock-solid reliability of its switches and networks. Telcos are control freaks, and they measure the performance of their systems out to five decimal points.

Don Listwin, who runs the Cisco unit that markets to telcos, says having a network that runs like a Swiss watch is only part of what these giants want: "Reliability is just table stakes. It in no way translates into an ability to innovate rapidly and bring new services onboard." That's where Cisco has much more experience than Lucent, he argues--it has been the key player in the breakneck evolution of corporate data networks. What's more, the tight relations Cisco has built with corporate CIOs will help it command the telcos' attention. Buy gear from Cisco, the pitch goes, and Cisco will steer those loyal CIOs to you when they want to outsource their data communications.

Cisco's ability to leverage its marketing power may be what worries Lucent most. McGinn says, "I tip my hat to them for having the most aggressive selling machine in the industry. They're nonpareil." By contrast, he says, the network systems group he once ran at AT&T focused on making the product, not selling it: "We would announce a product and not have a field force ready to do anything with it. When Cisco announces a product, their field force is unbelievably ready."

McGinn has made getting new products to market quickly Lucent's top priority. For starters, he's put a blowtorch to the undersides of his sales force, completely revamping the organization. He has also created tighter links between various Lucent operations. Researchers at Bell Labs, for example, now take extended tours of duty with the divisions that build the products they invent. McGinn's goal is simple, but it would have been considered radical at the old AT&T: release products according to schedule, even if keeping them in the Labs longer would make them a little bit better.

It's a different pace for Lucent, one measured in heartbeats rather than months. Bell Labs President Dan Stanzione shakes his head ruefully as a reporter starts to ask about the old rap against Bell Labs. "Yeah," he says. "Great technology, too slow." His key task in the battle with Cisco is to develop switches for the data-communications market that won't kill off services that the phone companies spent billions establishing, like 800 numbers and caller ID. Such services depend on software code that Lucent has buried in today's telephone network--software that took an astonishing 10,000 man-years to create. Moving all of that onto the new data network is enormously difficult, and enormously important. Says Stanzione: "If we can move features buried in voice switching, and add data services like voice over the Internet, we will win." So Stanzione has cloistered his best software engineers in a room in Chicago, to find a way either to dump the code directly from digital phone switches into computers that will control the new class of switch, or to create a way of automatically translating the old code into something a new machine could understand. This time, the engineers don't have 10,000 man-years.

That's the new hard reality at Bell Labs. The first team assembled to develop the routing switch (or is it the switched router?) said the job would take two years. Told that senior management would really like it in one year, they went back to the drawing board and said they could shave two months off the schedule. The team was promptly disbanded and told to find work elsewhere. The next group said they could do it in a year. They're set to finish work on schedule.

For a self-professed adrenaline junkie like McGinn, this sort of pressure is fun. Smart, fast-talking, and amusing, McGinn has an easy manner and an engaging curiosity. A native of New Jersey, he was an unlikely candidate for a career in telecom, having gone to Grinnell College in Iowa to study history. But he loved business and sent a resume to Illinois Bell after his senior year, winning a job in the sales department. He didn't plan to stay long, and it appeared he wouldn't have to; on his very first sales call, he was offered a job by his customer, an experience that was repeated often over the years. But every time he thought about moving on, he got a promotion. He has worked for AT&T and its spinoff for 29 years.

His current job is the culmination of a long climb back from a brush with the most ignominious episode in AT&T's recent history. After joining its computer division in the late 1980s, McGinn became president in 1990, just four months before AT&T launched its disastrous takeover of computer maker NCR. McGinn, who says he was surprised by the acquisition, had to disband his division and fold its remnants into the newly acquired company. When he finished that job, he landed at AT&T's network systems group. By 1994 he was running it, the biggest of the businesses AT&T decided the following year to spin out as Lucent.

In facing off against Cisco, McGinn knows that mutability is a prerequisite. For starters, he wants to expunge Bell Labs' not-invented-here bias. Says Schroder & Co. analyst Philip Sirlin: "They know they can't hope to develop everything internally." McGinn has already purchased several small companies that can beef up Lucent's data portfolio, including Yurie Systems, the company he just bought for $1 billion. (Yurie is a rising player in a switch technology called ATM, for asynchronous transfer mode, which is of great interest to phone companies because it can handle voice and video as easily as it moves data.) Bigger purchases may be in the offing: Up to now Lucent has been barred from using a low-tax merger method called pooling of interests because of some tax benefits associated with the spinoff, but that prohibition expires Sept. 30. One possible target is Ascend Communications, the $1.2-billion-a-year Alameda, Calif., company that is the industry leader in ATM.

However quickly it can crack the data market, the verdict on Lucent is, at the very least, so far, so good. At the time of the spinoff, the businesses Lucent inherited from AT&T were growing roughly 9% a year. In the last quarter (its second of fiscal 1998) Lucent's growth rate was 25%. Its market cap has increased from $19 billion to $97 billion. The buoyant sales of old-world telephone-network gear may well confound McGinn's prudent skepticism for a while to come, giving him some breathing room. As his colleague Bill O'Shea points out, there is more of the past in the future than most people think.

In the long run, the customers of both Lucent and Cisco may end up bringing the two competitors together, as AT&T has done already--buying data switches from Cisco and new fiber-optic transmission technology from Lucent. Until Lucent develops the moxie to play in Cisco's playground, it is happy to be king of a sandbox that still yields many treasures.