"Every hour, 3,000 miles of optic fibres are manufactured. That's enough to go around the world three times a day."
Bob Martin likes to talk technical, especially when the talk turns to what he calls "a global communications revolution". And so he ought. As chief technical officer of communications provider Lucent Technologies and vice-president of Bell Labs, the research and development engine of Lucent, he is the main technical man in an organisation with one of the most respected names over the last 30 years of computing and information technology.
Lucent spun off from AT&T three years ago, taking the prized reputation of Bell Labs with it, and in the meantime has risen to become one of the world's significant players in the communications equipment and software field. Its revenues last year were $30 billion (€27.5 billion), and although 70 per cent of this came from US business, it now has a presence in more than 90 countries, including a cable manufacturing plant in Bray employing around 500, and a customer services centre in Dublin.
But Mr Martin leaves talk of Lucent's business structures to others, preferring to analyse the technical trends that - depending on your opinion - are either leading or being led by the information and communications revolution.
He identifies five so-called "disruptive technologies", which in his opinion are fundamentally changing the marketplace: silicon, wireless, optics, packets, and software.
Silicon, he says, is still following Moore's Law (Gordon Moore, who in 1965 said the number of transistors on a silicon chip would double every 18 months). "But silicon is the slowest technology curve," says Mr Martin, pointing out that the speed, or bandwidth, of fibre optic cables is doubling every 12 months. Whereas each glass strand of a fibre optic cable once carried a single beam of light, Lucent this year expects to get 1,000 different wavelengths (colours) of light into a single fibre, multiplying the amount of information that can be sent through it correspondingly.
"The best commercial systems are 400 gigabits per second (by Lucent)," says Mr Martin. But he adds that research this year is aiming for 100 terabits per second on a single fibre. That's a million, million bits per second, or 17 million times faster than a standard 56 kbps modem.
But fibres are not the only way of communicating. Wireless technology has its own growth curve, with the number of voice calls per volume of air doubling every nine months, says Mr Martin. (The term wireless no longer applies just to old valve radios collecting dust on shelves, but is increasingly used to describe radio-based communications such as mobile telephony, MMDS and others.)
Mr Martin's fourth disruptive technology uses both fibre and wireless, and is called packet technology. This allows voice calls to be sent over data networks as many discrete packets of data, replacing the highly reliable but less efficient traditional circuit. Packets are a little cheaper: they are only sent when there is something to send, and people only talk for 40 per cent of the time while on a call.
But packet-based networks are less reliable and data travelling over them are subject to longer delays than on circuit-based networks. This is why voice calls over the Internet, for example, are of a poorer quality than traditional telephony.
Furthermore, putting voice into packets is not that simple, says Mr Martin. In the old voice networks, the predictable call pattern of users making three three-minute calls during peak hours meant bigger networks worked better. "The traffic statistically added up nicely," says Mr Martin. But in the data world, traffic is less predictable, he says, comparing it to traffic on large, busy motorways, where for no apparent reason a bubble of congestion may form.
"The bigger the network the uglier," he says, adding that service level agreements and new standards will be necessary to ensure adequate minimum quality for voice calls over data networks. Tying these four developing technologies together is software, "the glue" according to Mr Martin. The relatively slow development of silicon is leading to a software world, he says. Software is also more flexible, easier to change once installed compared to fibre optics, for example. When it comes to improvements in communications networks, he advises: "Put in the fibre first; the rest is software."
The convergence of voice and data is reflected in the convergence of the telecommunications and data communications business worlds, with a spate of mergers and acquisitions between firms supplying these once-distinct markets. In January, New Jersey-based Lucent bought Californian data communications provider Ascend for nearly $20 billion. Structured as a merger to take place later this year, this is intended to strengthen Lucent's challenge to world data communications leader, Cicso.
Other companies formerly in the telecommunications camp are doing likewise. French group, Alcatel, last week announced the acquisition of Californian data firm Xylan for $2 billion. Last August, Canadian firm Nortel paid $7 billion for Bay Networks, while industry analysts are expecting similar acquisitions by European and Japanese telecommunications firms.
German telecommunications group Siemens confirmed on Monday it was buying two US networking start-ups, Argon Networks and Castle Networks.
How is this industry convergence reflected in technology? Mr Martin says he sees three roads to convergence of voice and data. The first involves getting the circuit-switched networks to communicate with modern data networks, an approach favoured by the telecommunications providers, which have big investments in circuit technology. The second approach is typified by Internet service providers, who will add services to and improve the quality of their Internet-style networks.
Finally, in the heart of the networks, he says, faster communications over fibres and eventual improvements in the core protocol (communications rules written in software) which runs the Internet, called IP, will mean simpler and faster switches and routers along the way. Mr Martin's conversation contains an effortless array of mathematical algorithms and research results, all reinforcing one clear message: we're living in a communications revolution.
That point made, he says we should be trying to identify the services which will generate the data to fill all those radio interfaces and thousands of miles of fibre optic. It is time for entrepreneurs the world over to don their thinking hats.
Eoin Licken can be reached at elicken@irish-times.ie
