The speed of light

A little after 4 a.m. on Feb. 10, in a hotel suite in San Francisco, John Roth got a pleasing surprise. His angular 57-year-old body stubbornly functioning on eastern standard time as usual, the Nortel Networks Corp. CEO was wide awake, so he opened his door and retrieved The Wall Street Journal sitting outside. And in its pages he read that for the first time in his company's history, Nortel's market capitalization had surpassed that of Lucent Technologies Inc., the huge New Jersey-based telecom/fibre-optics conglomerate that has always been one of Nortel's biggest competitors. At one time that would have seemed impossible, a mistake, but there was no doubting the numbers: Nortel sat at $166.8 billion, Lucent at $161.8 billion. (All currency in U.S. dollars except as noted.)

Over Roth's incongruously medieval face--the visage more of a 12th-century woodcutter than a paragon of high tech--passed a flicker of satisfaction. This was the kicker to a very good week. On Feb. 2, the Chicago investment bank William Blair & Co. LLC, which tracks the high-tech sector and influences the decisions of some large U.S. institutional investors, had issued its latest report on the company, stating, "Nortel's opportunities for growth are significant, its strategy is visionary, its product portfolio is stronger than ever, and its competitive position is solid." And a few days after that, Roth had received an even louder huzzah at a Nortel sales conference, when some 4,000 of the company's salespeople had risen off their chairs in the San Antonio, Tex., convention centre and given him a standing ovation.

Since taking over from Jean Monty, now head of BCE Inc., in the fall of 1997, Roth has established himself as Canada's, and arguably the world's, leading proponent of the "high-performance Internet," the multilane version of today's electronic alleyways. He has made speeches in Europe and North America calling for "Five 9s" reliability, the 99.999% likelihood that, as with the telephone's dial tone, the Internet will be there when you need it. And he has made himself a voice of national change, publicly criticizing Canada's relative paucity of Internet IPOs, and making the brain drain issue his own, charging first that the country's tax structure, and then its health-care system, push its best high-tech talent south.

At the same time, he appears to have pushed every possible right button in his own corporation, turning it on a dime, it seems, from a lumbering old-style telecom supplier into a paradigm of performance, with 26% annual revenue growth, a more than fourfold increase in share price since the fall of 1998--so sharp that it skewed the valuation of 37%-holder BCE Inc., forcing a spinoff--and a clear global lead in leading-edge fibre-optics technology. "You'd like to be critical of your competitors," says John Chambers, CEO of California's Cisco Systems, still the world's leading manufacturer of Internet equipment. "Unfortunately I can't. I think John's done an excellent job at Nortel. He has developed into probably our toughest competitor."

John Roth, you could say, is the very model of a major modern success story, a darling of the market, an Internet provocateur and an exemplar to his peers, having pushed Nortel to a decisive win in this year's Most Respected Corporations survey. Yet there's a tinge of bitter irony in his nyuk nyuk laugh when he rises, casual and sweater-clad, from the elegant oval table in his Brampton, Ont., office and says, "All we had to do was drive the market cap to be 26% of the TSE and whaddaya know--you get respect." Roth is including BCE's soon-to-be-sold stake in his calculation, but the point's the same. If you're John Roth, it's not easy to forget that just a year and a half ago, a number of voices in the telecom and finance communities were calling for your head.

It was in August of 1998 that Roth completed the first stage of a process he'd begun upon taking over from Monty. Having announced in a Dec. 3, 1997, memo to employees that Nortel was going to make its now celebrated "Right-Angle Turn" toward Internet technology (from "dial tone" to "Web tone," he explained, a technology-mixing metaphor he's still proud of), Roth spent the following months looking for the right Web-tech company to buy. He went through what he called "the ABCs of the industry." He talked with Ascend Communications Inc., he spoke with Bay Networks Inc., he had discussions with Cabletron Systems Inc. and even Cisco. When he got to the Ns, he met with Terry Matthews at Newbridge Networks Corp.

He needed a company with expertise in IP, or Internet Protocol, the working language of the Web. He needed it because he knew that voice transmission, Nortel's traditional strength, was being overtaken by data transmission as the dominant form of network traffic. He knew that IP was riding the curl of a bigger wave than the competing data transmission protocol, ATM (the Betamax to IP's VHS), and he had to signal to the market, and to the "ATM bigots" inside and outside Nortel, that he was serious about the right-angle turn toward IP data gear. His deadline was even more pressing because he knew that Richard McGinn, CEO at Lucent--Nortel's U.S. counterpart after it was spun off from AT&T Corp. in 1996--planned to go shopping for the very same thing in September. If Roth didn't want to get into a bidding war, he had to buy now.

Bay Networks, based in Santa Clara, Calif., became the likeliest option because it offered IP ability in the arena of enterprise networking--the hardware that corporations use to communicate--which complemented Nortel's strengths in the long-haul core networks that run like superhighways between cities. And because Bay's revenues had been flat for a year, it was also affordable.

It took much of the spring to convince Bay's David House to go along with the idea. But in June, Roth announced the $6.5-billion deal, with 270-million Nortel shares and

78.8-million options headed to Bay shareholders. In August, the largest acquisition in Nortel's history was done.

By that time, however, it was no longer a $6.5-billion deal, it was a $4.4-billion deal, because the market had already decided that Roth had bought the wrong company, for too much, and begun to drag Nortel's stock price down. "Everybody," says Roth now, "thought I'd lost my marbles."

The market's confidence had also been rattled by a coincidence of events. In June, one of Nortel's best customers, Sprint, announced that it was going to stop buying Nortel switches and instead wanted Cisco to be the main supplier for a new ATM network, a move that some IP converts within the industry considered wrong-headed but which nevertheless set a dark little cloud over Nortel. At the same time, Roth was trimming away some "deadwood," laying off several thousand employees to reduce the total workforce to 75,000, which the Street read as a sign that earnings were weak. "We didn't explain it well," Roth admits. The market became increasingly skeptical about the company's ability to meet its ambitious growth targets. And then came the calamity of Sept. 29.

At Nortel's annual financial conference at New York's Essex House, Roth and Nortel CFO Wes Scott were on stage, walking hundreds of fund managers and analysts through the company's plans and projections. The audience was nervous: The Asian meltdown was well under way, and France's large telecom company Alcatel SA had just disappointed the market with terrible earnings, suggesting that the entire sector might be vulnerable. During the question-and-answer session, Scott let slip to this fidgety crowd that the numbers Nortel had suggested during the morning session might not, in fact, be immediately achievable.

What the heck does that mean? the managers wanted to know. Because Nortel was due to announce its real numbers in three weeks, it was in its official silent period, so Roth and Scott couldn't explain themselves. It appeared that someone in Nortel had been giving "double guidance"--negative advice to some analysts and encouragement to others. "The whole presentation kind of fell on its face," says Benoît Chotard of National Bank Financial. "Basically it's run for your cellphone--'Sell Nortel!'"

"We all sat up there looking very evasive," recalls Roth, "and the more evasive we looked the more the audience was sure we were hiding something, and everything just started to tailspin." Within a couple of days, Roth had watched Nortel's stock dive 20%, vaporizing $9.1 billion (Canadian) in its market capitalization.

In the stressful days that followed, did Roth go to Jean Monty for advice? "No," says Monty wryly. "When these things happen, many people on the board give advice freely without the CEO asking for it." Not immediately, but some months later, Scott and the head of Nortel's investor relations department were reassigned.

It wasn't one thing but a combination of factors--foresight, luck, sweat and engineering chutzpah--that has enabled Roth and his crew, over the past 18 months, to pull Nortel out of that puddle and into the stratosphere. What's certain is that, like a sprinter in his crouch, much of what Nortel needed to spring forward was already in place.

It had, first of all, Roth. Here was the son of a onetime ham-radio operator, a kid who'd grown up in Lethbridge, Alta., Red Lake, Ont., and Winnipeg, watching his father construct hulking transmitters and receivers. Electrical engineering was in his genes. He got his master's in engineering at McGill University, then worked for RCA, designing antenna systems for the Canadian navy and ground stations used to communicate with the earliest satellites.

When he joined Northern Electric Co. Ltd. (later Northern Telecom Ltd.) in 1969, Roth devised an innovative design for what would become the Anik satellite, shaping it to approximate the contours of Canada, as seen from orbit, to focus its signal. Northern Electric lost the contract to Hughes Aircraft, but the final design was close to the one Roth had conceived.

When he headed a team in a new digital radio effort in the mid-'70s, Roth realized that leading the project and making decisions about what would be designed was more fun for him, and more significant for the company, than designing the product itself. He took on a series of managerial positions until, in 1982, at age 39, he was made CEO of Bell-Northern Research Ltd., Nortel's research division. And he was still Nortel's R&D chief when Paul Stern arrived as CEO in 1989.

Stern's leadership, during which he cut R&D spending from the traditional 14% to 11% or 12%, must have been galling to Roth. A business acquaintance once heard him admit that there were few things in life worse than being ripped by Stern at a board meeting. But today, Roth is politic. "I learned from Paul," says Roth, "what determination really looks like."

That Roth survived and even prospered--launching Nortel's wireless effort in 1991--during the regime of a man so devoted to short-term gain says a good deal about his ability to manage life's vicissitudes. "John," says a former veteran Nortel executive, "is a very adaptive individual."

By 1994, Jean Monty, brought in two years earlier to reconstruct Nortel's reputation and push its global effort, started handing Roth more and more of the company to run. And three years later, as the surge in data traffic was pushing telecommunications into a period of profound technical change, he was positioned to be Nortel's first CEO since Walter Light in 1979 to be intimate with the company's technological capability. "He had one other thing as well," says Jean Monty. "He had business acumen. Not all technology people understand the intricacies of business transactions or strategy. And he had that in his makeup."

Nortel had another muscle ready to fire. Roth was about to demand change from a culture already accustomed to transformation. The Consent Decree in 1956, in which AT&T was obliged by the U.S. Justice Department to sell off its foreign holdings, forced what was then Northern Electric to fend for itself, without access to the patents and technologies of AT&T's Bell Labs. In the 1970s, executives such as John Lobb and Walter Light jerked Northern Electric through a series of reforms that led in 1976 to a name change--to Northern Telecom--and a groundbreaking move into digital switching, a shift that came to be known as "Digital World."

"Roughly 20 years ago what you had was a brand-new company--new culture, new management, new approach, new structure," says Francis McInerney, a telecommunications analyst and principal at New York's North River Ventures Inc. "Call it the culture of the makeover. The culture of rebuilding was established."

Jean Monty capitalized on that culture when he took over Nortel in 1992. A year earlier, the company had echoed its '76 digital transformation with the launch of "Fibre World." But Monty saw that the company had become "complacent" and hadn't committed itself to catching the new technological wave. So the Montrealer established the Right-Angle Turn precedent by imposing what he called a virage, a sharp change in direction, toward a more international, more customer-focused organization. Nortel recommitted itself to exploring the high-capacity optical technology that would enable it to capitalize on the coming explosion in bandwidth demand.

Roth contributed to the culture of change when, as North American president in 1994 and COO in '95, he replaced the company's geographical organization with one defined around product groups such as carrier (its core network business), wireless and enterprise. It saved money and made Nortel more responsive, though at the cost of executives such as Gerry Butters, head of the United States unit, who left for the open arms of Lucent.

This reorganization, and the one that followed when Roth took over as CEO--aligning Nortel's efforts more around types of customer service than products--also let Roth salt the upper ranks with executives who saw the future as he did. Says McInerney, "He managed to basically eliminate anyone who was not loyal to him." Or to his vision.

"There are people who just aren't going to get it, and you gotta make changes," says Roth, who tends to speak as quickly and unaffectedly as he combs his hair.

Nortel may have needed a push into Internet technologies, but once jostled, it understood the process of reform. With the purchase of Bay Networks, and the decision to overlay the more responsive culture of its 7,500 employees onto the 75,000 at Nortel, Roth intensified the effort to remove the bits of constricting sinew from Nortel's bureaucracy. He made Bay's David House president, in part because he wanted House to have the authority to instill changes. But the two apparently chafed, and House was reportedly paid $4.5 million, not including options, to leave at the end of a year. Still, old Nortel practices such as "banding," which categorized management into levels that determined their pay and bonuses as well as their eligibility for promotion, are now being stripped away. Roth started with the top 500 executives. "I can't figure out whether you're this kind of executive or that kind of executive," Roth told them. "You're just called 'executives.' And this is what I expect of executives...leadership, passion, vision, determination to win, make decisions, get on with it. And if you don't have it, then you're culled."

The pace of change in the industry demanded ever more intense focus on the needs of customers, the ability not just to respond, but to anticipate. "What are you doing for the customer?" Roth asks, directly and in his quarterly e-mails. "How does your job help Nortel serve the customer? How does your job impact the customer? If you can't describe it, then maybe you should look for a different job inside Nortel."

But to gain that focus, he had to give people measurable goals, so the bonus system was changed in favour of the toughest, riskiest projects, the ones that might push Nortel out front. The company also had to find a way to keep its most valuable employees, who are constantly being called by high-tech start-ups in Silicon valleys north and south and tempted with potentially limitless riches. (Last October, Nortel sought a court injunction against Optical Networks Inc., partly owned by Cisco, to stem the raiding of its fibre-optics talent.) To resist that tug, and to spark a venture capital hunger inside the company, Nortel instituted a system of phantom stocks. A project is assigned an artificial IPO date. The sooner a product gets to market, and the more customer needs it satisfies, the more revenues it generates, which allows a healthier "IPO" at which employees would cash in, theoretically up to six or seven figures. Total payouts, Roth says, have been in the tens of millions.

"And don't talk to me about schedule slippages," Roth told them. "I don't want to hear about it. You just tell me what your revenue's gonna be. And if you get to market early, revenues'll be bigger, you'll be selling for longer. And if you're late, you might not have any revenue. If you have no revenue, there is no IPO." In that way, then, the program doesn't quite mirror the new market reality. But it's close enough to scare a lot of Nortel employees. Those who prefer security are given the chance to choose a more traditional project, and so the system self-selects the hungriest participants.

In every move Roth has made since becoming CEO, his emphasis has been on speed. He had good reason. "The biggest problem Nortel has had," a telecommunications analyst told The Wall Street Journal last year, "is not that they don't have good technology, they've just been late to deliver." As a maker of large, expensive hardware, Nortel was used to working in "long-long" time--taking years to develop products that would then last years. An employee opinion survey in April '99 suggested that many Nortel employees got Roth's hurry-up message and had an idea of how to improve the situation: Reduce the micromanagement and the politics among senior ranks. It was no surprise, then, that Roth's Right-Angle Turn memos--coming from a long-term Nortel employee who knew how things worked--went directly to employees, bypassing traditional communication lines and pressuring middle management to hurry up and get it or vacate the lane.

To further disrupt the company's inertia, Roth--his own garage full of fast vehicles with cornering power, including a Porsche 911, a '67 Corvette, a replica Shelby Cobra and a bright red Prowler--began running over projects that had been in development for too long. "There were significant pools of blood," he admits, with some satisfaction. Even Nortel's concept of time was changed. Roth noted that, as microprocessor speed has doubled every 18 months in the computer industry (known as Moore's Law), the capacity of fibre optics now doubles every nine months. (Within Lucent, this is known as Butters' Law, after the former Nortel exec.) Insisting that a year was therefore too long a time frame for project planning, he began talking about "Web years" (a fiscal quarter).

Rather than wait for Nortel's own R&D spending--whose Internet share has gone from 5% in 1997 to 60% in '99--to produce results, Roth has bought innovation by acquiring eight more companies after Bay Networks, including Cambrian Systems, Shasta Networks, Clarify and Qtera (see Roth Goes Shopping, page 38). Many of them were discovered early in their development because Nortel had put seed money into venture capital funds such as Accel, Battery, Redpoint, NEA and VantagePoint in the U.S., and Ascend in Israel, and so had advance warning of groundbreaking work.

Better that Roth get those companies than Cisco, which had been vacuuming up companies at the rate of 10 a year. It didn't matter what their numbers were. When Cisco acquired the revenueless Cerent, for about $7 billion, Roth thought the market would punish it. "What happened during the day is Cisco's stock went up, mine went down," he exclaims. "I thought, 'I bought the company for Cisco.' And that's the new world of acquisitions. That's the new math."

Corporate systems or cultures are only as good as the products and the market share they produce, of course. Today Nortel has hundreds of products competing in the hotly contested areas of carrier and enterprise (Accelar PBX boxes in the company basement, BayStacks in the wiring closet), high-speed wireless (GPRS, or general packet radio service), Internet access (where the throughway meets the on-ramp--1-Meg modems to speed home connections) and Internet telephony. In 1999, Nortel's wireless and enterprise divisions, measured against the likes of Cisco and Lucent, were not considered its strength. But the market hardly seemed to mind, because right now much of the heat in telecommunications is in high-capacity fibre optics. And here, says an analyst, "They're winning. They're winning big."

Roth sums up what Nortel does in optics this way: "We don't do the glass, but we light it up." At the moment, Nortel is lighting the glass brighter than anyone else. With products such as the OC-192--a wardrobe-sized box equipped with $60,000 slide-out cartridges (or "blades") that send data through laser units the size of a pistachio at 10-billion bits a second--Nortel is setting the standard for stuffing maximum data through the fibres of the Internet.

In the company's fibre-optics research headquarters in Ottawa, Greg Mumford, Nortel's head of optical networks, walks and talks with barely concealed conqueror's pride--his just-booted laptop even announces his arrival in the room with a perfectly timed Windows crescendo. He has reason to be chuffed. Nortel anticipated the explosion in demand for bandwidth, and exploited it first. It introduced its 10-gigabit product, the OC-192, when the world still thought 21/2 gigabits was the most it would ever need. In 1997, says Mumford, "Lucent was telling the world this was impossible, and we were shipping them out the back door by the hundreds." As a result, Nortel has emerged as the clear leader, with 26% of the global market in both SONET and SDH optical systems--the hardware that lights up the fibre according to North American (SONET) and European (SDH) parameters--and 32% of the global market in DWDM systems (dense wave division multiplexing, the means by which different light waves, each carrying a separate stream of data, are closely combined along one fibre). Lucent, despite being one of the companies that actually makes the fibre, comes second in both.

Not that Nortel didn't have some queasy moments. Early on, the market didn't want what Nortel had to sell. Even the new carriers--companies such as Qwest Communications International Inc. and Global Crossing Ltd., which were competing with the older telcos and should have understood the need for bandwidth capacity--questioned the need for a 10-gigabit laser. Roth and his people went to Qwest's CEO, Joe Naccio, and made an innovative offer. Nortel would install the 10-gigabit system at the price of the 21/2 offered by its competitors, with just enough blades to run at the 21/2 rate. Whenever Qwest needed more capacity, they'd come and equip it and charge for the upgrade. With no risk to him, Naccio went along with it.

"Couple of interesting things happened," says Roth. "He had a forecast of how fast he'd grow, and we had a forecast of how fast he'd grow. He outgrew both." Having more capacity to sell, Qwest had sold more than it thought it ever could. And now the industry wants all the capacity it can get.

With its DWDM technology to manage the wavelengths, essentially colours, of many finely tuned lasers, Nortel is now testing the limits of fibre's physical capabilities by transmitting 1.6 terabits of data (1.6 trillion bits--equivalent to 360,000 high-resolution feature films) in a test with giant carrier MCI WorldCom. And while Lucent and Cisco scramble to catch up to that number, Nortel has announced its intention to be the first to combine 80-gigabit lasers for a total of 6.4 terabits.

The point of all this data compression is to push the cost of data communication down. Voice pays far more--one of the reasons some telcos are slow to change--but its growth prospects are limited. Data pays a trifle, but its potential is infinite. "Guys who've got an expensive network can't target Internet traffic," says Roth. "They'll die." Nortel's purchase of Qtera, a company in Boca Raton, Fla., whose new technology could allow laser signals to be sent thousands of miles over long-haul fibre networks without expensive regenerators, should eventually mean huge cost savings. And he who sets the lowest cost wins. Which is why these days there's nothing sexier than a cheap terabit.

All of the vocal effort Roth has put into revving up his people has paid off in marketing terms too. It's no accident that Nortel's president of enterprise solutions, Bill Conner, who encouraged Nortel's Bay purchase, was named marketer of the year by a New York-based marketing magazine. Roth's personal memos to his employees and all the gnashing about achieving Web speed have become key elements in his company's campaign not just to change, but to be seen to have changed. In March of last year, Nortel launched its first mass-media advertising campaign, under the Beatles theme Come Together, and the next month the company came together under its new Nortel Networks Corp. brand. In September, positioning itself as both in touch with customers and in charge of the Web's evolution, it launched its multimedia What do you want the Internet to be? campaign. In October, at the industry's big event, Telecom '99, Roth pushed the concept of the "high-performance Internet," and on Jan. 30 of this year, Nortel predicted a Net economy of $2.8 trillion by 2003 in front of the World Economic Forum at Davos.

"He's completely reversed the public perception [of Nortel]" says North River Ventures analyst Francis McInerney of Roth. "And he's had to do it by training, leading his own people. I ask myself when I look at his background, where in hell did he get that talent from, 'cause it isn't in the résumé. That's an achievement."

Roth has done it with an outwardly genial, unassuming style that contrasts with the typical approach of today's high-tech executives. "He's not that charismatic Mr. Cisco," says National Bank Financial's Benoît Chotard. "Certainly from a personal perspective, he doesn't bring out the 'Gee whiz, this is a leader,' in the traditional American sense. But he's certainly an incredible strategist and certainly a very forward-thinking technologist. And that's what you need at Nortel."

But Roth's personal manner isn't without its critics. While admiring the way Roth seeks consensus around his ideas, a former long-time Nortel executive suggests, "He tends to approach problems in a very analytical and objective manner. And therefore tends to dissect them almost mathematically. As opposed to someone who would look at a problem more humanistically. ...That technique works best when you're dealing with a product or a specific market problem. It tends to work less well when you're dealing with a human-resources issue."

Rather than seek out the differing views of his executives, Roth tinkers on his own until he finds the solution to a problem, not unlike the engineer he once was, or the shade-tree mechanic he still is. He refers to this tendency himself when he talks of trying to replace Nortel's traditional method of negotiating executive incentive targets, a process he considered "really dumb," with something more conducive to productivity. He left on a three-week vacation, he recalls--"and it's dangerous because I think about things"--after which he promptly instituted a plan for non-negotiable targets. Even the all-important Right-Angle Turn, Roth says, came to him in the summer of 1997, when he went off seeking inspiration in Silicon Valley and "came to realize just how profound this Internet movement was."

The amount of attention paid to Roth's instigator role suggests that it's part of Nortel's marketing plan to have its CEO personify its forward charge. But Roth himself clearly has a profound sense of his own contribution. "The company has split its stock four times over its history," he told the Ivey Business Journal last fall, "and I've split it two of those times."

"I think that John can sometimes come across as a little arrogant," says his former colleague, who also notes with alarm that Roth has stated publicly he's fully prepared to take on Cisco Systems. "To taunt your competition has no upside."

But there's no question the rivalry is intense. Last year, Nortel staved off Cisco's poaching of its Telia phone system business when Cisco wasn't ready with the Internet voice-data equipment the Swedish company wanted. "Nortel made a big deal out of the fact it was not being replaced, which was charming," sniffed Larry Lang, vice-president of Cisco's service provider division, to an industry publication. And Sprint, which had once seemed so ready to move to Cisco's data technology, has in the past year given Nortel $520-million worth of new business.

Now, not only is Nortel trying to cement its advantage in fibre-optics networking by piling $260 million into its operations in Ottawa, Montreal and Northern Ireland (in addition to the $450-million Ottawa/Montreal expansion already under way), it is also attacking Cisco where it lives--in the routers that direct transmission traffic.

Cisco has more than 80% of the router market, an insurmountable lead. So rather than sweat for a meagre number of router sales, Nortel's strategy is to make routers essentially valueless. Much as Sun Microsystems Inc. has done with Java, Nortel has licensed out the use of its own routing software to some 100 companies, including Intel Corp., Microsoft Corp. and Motorola Inc. The idea, which Nortel calls "Open IP," is to reduce routing from a proprietary product to a mere function, found on a microchip in a cellphone, a Palm Pilot, a microwave oven or any number of appliances that could soon be connected to the Internet.

It's all part of the relentless trend dragging the cost of information toward nothing, where only the fastest, lightest companies will survive. And no matter how much Nortel has changed, Roth must face the fact that the industry paradigm is still defined by Cisco, a company whose internal systems are so swift and unconstrained that one industry observer describes it as a soccer ball--completely hollow.

The day before Roth read the happy news that his company, after all those transformations since 1956, had finally surpassed Lucent in market value, another bit of news surfaced, which The Wall Street Journal barely bothered to report. Cisco Systems, just 15 years old, had briefly moved ahead of General Electric Co. with a market value of more than $480 billion, which for a time made it the second-most valuable company in the world.

And at the end of an interview, when John Roth looms over the Nortel share price displayed on his computer's Yahoo! screen and whistles, "So we're back up!," you know that's the Web speed that really matters.

Roth Goes Shopping

Before the Internet revolution, Nortel could afford to look disdainfully on the products and processes created outside its own R&D labs. But with the pace of innovation accelerating, and with chief rival Cisco Systems collecting new companies at the rate of 10 a year, Nortel CEO John Roth can no longer wait for inspiration to smite his designers. Since taking over in the fall of 1997, Roth has not hesitated to acquire what he needs to fill divots in Nortel's portfolio. APTIS When: April 22, 1998. How much: $305 million (all funds in U.S. dollars). Why: It offered network access switches to help carriers and Internet service providers meet the growing Internet demand, and smoothed Nortel's path from the world of voice to data networking.

BAY NETWORKS When: Aug. 31, 1998. How much: $4.4 billion. Why: Bay, a maker of IP-based gear for local area networks, was just what Nortel needed to announce itself as a serious data-networking player. It was also what archrival Lucent needed--in fact, most analysts thought a Bay-Lucent merger was likely--so getting it was a double coup.

CAMBRIAN SYSTEMS When: Dec. 15, 1998. How much: $300 million. Why: It's from Cambrian's DWDM (dense wave division multiplexing) technology that Nortel got its OPTera Metro product--the fibre-optic "Internet on-ramp" hardware that carries up to 160 gigabits of data per second

for enterprise (corporate network) customers in metropolitan areas.

SHASTA NETWORKS When: April 16, 1999. How much: $340 million. Why: IP services.

The Sunnyvale, Calif., company builds software tools to give carriers

the ability to create things like network-based firewalls, virtual private networks and e-commerce without having to add expensive hardware.

PERIPHONICS When: Nov. 12, 1999. How much: $378 million. Why: Nortel reportedly plans, though has yet to announce, the creation of a $1-billion software division to help its customers handle their e-business customers and applications, and Periphonics offered voice-enabled call-centre software, allowing Nortel to bill itself as a full-service, end-to-end supplier.

CLARIFY When: Announced Oct. 18, 1999. How much: $2.1 billion. Why: Clarify makes customer relationship management (CRM) software, letting companies track and serve customer needs. Adding Clarify to the call-centre software provided by Periphonics addresses Nortel's perceived weakness in CRM.

PROMATORY When: Announced Jan. 6, 2000. How much: $778 million. Why: Nortel had a non-exclusive deal to use Promatory's IMAS Internet-access product, which enables a single enterprise phone line to handle multiple transmissions of voice and data. Then Nortel decided it didn't want to share.

QTERA When: Completed Jan. 28, 2000. How much: $3.25 billion in stock. Why: At the moment, data can't be sent more than about 500 kilometres over optical fibre before it starts to erode and must be regenerated with costly equipment. But Qtera has the technology to power a data transmission along a fibre for 4,000 km. Ipso facto, a lower-cost network.

DIMENSION ENTERPRISES When: Completed Feb. 9, 2000. How much: $64.5 million.

Why: Nortel needed to beef up its e-business consulting side. With its

skills in designing Internet data centres and integrating IP networks

with revenue-generating applications, the Virginia-based DEI, with clients such as AT&T, gives Nortel the supper to put on its stove.

xros When: Announced March 14, 2000. How much: $3.2 billion in stock.. Why: It developed an optical switch that uses tiny mirrors to redirect light waves without first having to convert them to electronic signals, which competes directly with a new Lucent device and supports Nortel's bid to build an all-optical network.