John Roth is worried his company may be getting too big. With revenues growing at about 40 per cent annually, he says Nortel Networks could soon become too large and complex to manage. It is a nice problem to have - and one that is largely of his own making.

In the three years since he took the helm of the Canadian telecommunications equipment maker, Mr Roth has transformed Nortel into a networking powerhouse that can barely keep up with demand for its hardware, which carries huge amounts of traffic across the Internet. Nortel currently employs more than 2,500 people at its plants in Belfast, Galway, Dublin and Shannon. Up to 2,000 more jobs have been announced for these factories.

Mr Roth has become something of a darling on Wall Street and on Bay Street in Toronto. Analysts praise his business plan to reinvent the company and his adeptness at executing his strategy. Investors also love the engineer-turned-chief executive who, after briefly stumbling in late 1998, has overseen an eight-fold rise in the value of Nortel shares, giving the company a market capitalisation of $235 billion.

Mr Roth's love of engineering originated with his father's ham radio. He says he once gave a school science class speech about ham radios. He recalls being fascinated by how they worked and how to build them.

This drive to understand how things worked extended to automobiles, a passion he still has today. But he was worried that he might end up designing washers and brackets rather than autos, so he decided to pursue his interest in electronics.

Mr Roth graduated from university in the mid-1960s during one of the most exciting periods of space exploration, when satellites were being put into orbit and the US was striving to reach the moon.

In his first job at RCA, he helped design ground stations for satellite systems. "The fascination with space communications and the promise it brought was just enthralling," he says.

The years have clearly not diminished Mr Roth's sense of excitement.

He says engineers like himself are driven to build monuments to themselves by developing products that are used and valued by society. At RCA he learned to question the usefulness of projects he was working on. He often could not identify customers who might buy the products he was being asked to develop.

"And so I became a little different, because most people were content, if the project was funded, to work on it. I was not. I had to know there was a customer . . . because I wanted to see this thing in use," he says.

Mr Roth moved to Nortel in 1969 and eventually moved out of the labs because he wanted a say in the projects the company would pursue.

"I'm an engineer who pays a lot of attention to the market. Instead of reading Scientific American I read Business Week and look at market trends, which, you know, more engineers need to do," he says.

But he was not always so successful. His initial foray into the data world appeared doomed to fail. No one quite knew what to make of Nortel's $7 billion acquisition in 1998 of Bay Networks, the data networking group struggling to keep up with Cisco Systems.

Most observers did not like it and when, in late 1998, it appeared that Nortel's traditional voice-based equipment sales would suffer as a result, investors punished the company.

Mr Roth says he never doubted his strategy to transform Nortel so it would be well positioned to meet demand for high-speed data networks based on fibre-optic technology. With 17 acquisitions worth more than $30 billion under its belt, Nortel is now widely seen as a leading contender in the race to develop the next-generation wireless Internet.

Personable and down-to-earth, Mr Roth likes to recall how he was sitting at home surfing the Internet one day in 1997, looking for a part he needed for one of his vintage cars. He eventually found the item via the Internet at a shop in north London. He refers to this episode as his Internet "epiphany" - the moment he realised the Internet had the potential to become a great equaliser.

"Its potential for business was certainly not obvious to a company like Nortel, which viewed the Internet as a threat because we were building voice networks."

Forcing Nortel to see what is now obvious was not easy. Soon after his appointment in 1997, he sent a memo to employees outlining the company's new strategy. The challenge for Nortel, he said, was to build robust, reliable Internet-protocol networks, which could carry massive amounts of data more efficiently and overcome a tendency to crash. Nortel, he said, must make these networks as reliable as the voice networks on which it had built its reputation.

After snapping up Bay Networks, Mr Roth was determined to make Nortel a nimble competitor that could get products to market quickly.

Mr Ken Leon, an analyst at ABN Amro, says Mr Roth has brought quiet, steady and strong leadership to Nortel at a time of significant upheaval. Mr Roth is seen to combine the speed and flexibility of an Internet player with the gravitas of a chief executive of a large telecoms equipment maker.

"He can smell opportunities and he is extremely strong at executing," says Mr Leon.

It is not the first time Mr Roth has spurred Nortel into changing its ways. He is also widely credited with having steered the company into the wireless sector after European rivals stole a march on their North American competitors. As manager of Nortel's research and development budget in the late 1980s, he says it was not easy persuading his superiors to focus on wireless and data communications. These two segments now account for a significant share of Nortel's revenues.

Mr Roth recently handed over day-to-day operations to Mr Clarence Chandron, Nortel's new chief operating officer. That has given him more time to focus on the larger strategic issues facing the company. Among the more pressing is to identify new opportunities for the company and to prevent it from being surpassed by any big technological development.

Mr Roth is also busy working out how to keep control of his rapidly expanding empire. Nortel's growth forecasts suggest the company will generate $40 billion in sales next year, reaching a scale and complexity of operation that Mr Roth believes could become unwieldy if not properly managed.

He is well aware that many high-flying companies have faltered in the past because management teams have been unable to keep pace with rapid expansion.

"The problem is that we are growing extremely quickly. Can our processes and management strength keep up with the size we are becoming?" he asks. "We're putting a lot of energy into simplifying Nortel so we can still run it when we break through the $40 billion barrier."

As challenging as it has been to bring Nortel this far, Mr Roth recognises that successful expansion beyond the $40 billion threshold will depend on how he himself develops as a manager and leader.