Life at 100 Billion Bits Per Second
Fiber optics is going to make communication so cheap
that we will be spending much more money on it than
we do today.
The law of the photon
By Howard Banks
MOORE'S LAW said that chip power would double
every 18 months. That's plodding. The new law of the
photon says that bandwidth triples every year.
Have you ever given up on the Internet because you
got bored waiting for a photograph to compose itself on
your screen? Your modem is not the main villain. Your
lack of bandwidth is. Even with the fastest modem in
the world, if traffic is heavy that image is going to take
forever to arrive.
Bandwidth is the carrying capacity of a communications
line. It tells you whether your telephone line is just
good enough for a plain old telephone call-or can give
you movies on demand, teleconferencing, remote
diagnostics, everything you ever wanted from the
Internet with no wait, things that you can't even
imagine today.
Triple every year? It was a wild claim when silicon
commentator George Gilder first made it in 1993.
Geometric growth at that rate is rarely seen in human
activity. If anything keeps up that growth pace, it grows
a billionfold in 19 years. Does anything grow that fast?
Fiber optics comes close. Experimenters at Lucent
Technologies' Bell Labs have pushed the speed limits in
the laboratory up by a factor of nearly ten in the past
two years, to 3 trillion bits-3 terabits-per second. At
the 28.8-kilobit transmission speed of garden-variety
modems, that's enough for 100 million simultaneous
Internet connections.
In the space of two years MCI has raised the bandwidth
of its Internet backbone by a factor of 8, to 1.2 gigabits
per second. But for a country where a million homes
want to see video on Web sites and movies on demand,
1.2 gigabits won't cut it. For them, those multiterabit
connections now on laboratory benches at Lucent and
elsewhere will be necessary. When will terabit lines be
available? In not more than five years.
Making all this possible is photonics, the science of
sending data bits down pulses of light carried on
hair-thin glass fibers. There is no official name for the
law that says how fast this science will carry us into the
next century. We could, however, call it Payne's Law, in
honor of David Payne, a 53-year-old physicist at
Britain's University of Southampton. Payne is perhaps
the leading scientist behind two key inventions in
photonics over the past decade and a half.
Significantly, both can be retrofitted onto fiber already
buried in the ground.
One is the optical fiber amplifier, an ingenious device
that makes it possible to magnify the reach of a light
pulse without first converting that light to electrical
pulses and then back into light. That amplifier is vital.
Without it, photonics would be advancing, but we
wouldn't be seeing any tripling of power every year.
Payne's second major contribution is an enhancement
to the amplifier that corrects for the distortions in light
pulses-a smearing of the image, so to speak-that
occur when an optical signal is pushed to the limits in
bit speed and distance between amplifiers.
At the same time that Payne was leading the way with
these two inventions, other scientists were advancing a
third technology for expanding the capacity of optical
fibers. It is called wave-division multiplexing. In plain
English, it is a method for simultaneously dispatching
laser pulses of different hues down the same tiny fiber.
Credit goes to hundreds of scientists at half a dozen
firms, including Lucent, the Italian tire- and cablemaker
Pirelli, Corning Glass and Ciena (see story, p. 70).
The world in general has yet to appreciate the impact
this science will have on our daily lives. "We are really
only in the Stone Age of optical communication," says
Professor Payne. William Gartner, Lucent's vice
president for optical networking products, talks about
the possibilities. "For businesses and consumers,
applications will emerge that today we don't even dream
about. High-speed Internet access and video
interconnecting all homes will be a reality, there's no
question of that.
"People are exploring things like remote surgery today.
The need for bandwidth is just dramatic there. Optics
will allow networking of huge bandwidth from anywhere
to anywhere, so it's maybe the Mayo Clinic tied in with
NYU, tied in with the University of Houston, all
collaborating on this patient who's being operated on in
Argentina. Doctors don't even fathom that today."
But they will soon. "Progress [in opto-electronics] is
faster even than microchips were at the equivalent point
in their development," says Gerry Butters, president of
Lucent Technologies for the North American region.
Electronicast, a San Mateo, Calif. market analyst, says
that sales of opto-electronic equipment hit $4.5 billion in
1996 and will grow to $34 billion by 2006.
After that? The sky's the limit. Two avenues of current
research in optoelectronics could make the next 15
years as momentous for communications science as the
past 15. One is optical switching. If amplifiers can be
purely optical, why not switches, the computers that
route all those phone calls and all that Internet traffic
among hundreds of millions of endpoints?
Everything carried on optical fiber, whether it's a phone
call, a data file or video, starts out as electrical impulses.
Before they can enter the fiber, they have to be
converted to optical form. Today that's done at the local
phone company office, using a costly computer-or
switch-that modulates a laser so that variations in the
light carry the signal (see diagram, p. 72). The difficulty
is separating those different messages to deliver them
to their ultimate destination.
An optical switch would make the transfer more reliable
and cheaper. Following a successful demonstration
program led by Darpa (the Defense Advanced Research
Projects Agency) in late August, Hitachi Telecom
(USA) announced plans for a commercial trial of an
optical switching, or cross connect, system on MCI's
optical network in the Dallas area. Lucent's Butters
reckons that all-optical devices allowing signals to be
cross-connected should be commercially available by
1999.
The other glittering goal is so-called digital optics.
Scientists at British Telecom and in many other labs
around the world are looking for a way to manipulate
light pulses as nimbly as they manipulate the voltages
of a transistor circuit. "What's needed," says Payne, "is
the equivalent of the electronic world's ability to take in
a weak signal, reshape it to its precise original form,
eliminating unwanted 'noise,' and then reamplify it and
send it on its way. Optics as yet can't do this
reshaping."
Terabit backbones, optical switches, digital optics. Stir
them together, allow a decade or so for development
and we arrive at a stage where communication will be
priced in microcents per minute. But this is a market that
is highly price-sensitive: The cheaper it gets, the more
of it people will use. So what if your phone bill doubles
or quadruples, so long as the extra money brings you
first-run movies and all kinds of pleasures and
conveniences?
As competition grows in world telecommunications and
national boundaries fall, Payne predicts, there will be no
cost difference between a call around the world and one
to the corner grocery shop. By today's standards,
communication will be ridiculously cheap. Which is
precisely why we will be spending more money on it
than we do today-and why the telecom business can
only grow and grow and grow. |
