Future Boy
What Ever Happened to Broadband?
By Erick Schonfeld
business2.com
The bankruptcies get announced with depressing regularity these days:
Teligent, 360networks, Covad Communications, Exodus. The losses keep
piling up: $55 billion for JDS Uniphase alone; more than $20 billion for
Nortel Networks. And the stocks -- once the envy of the modern world --
well, don't bother looking unless you have a strong stomach.
Telecommunications companies are going through a harrowing transition
right now. The old strategies, predicated on unlimited demand for
bandwidth and unlimited access to capital, no longer apply. Companies in
every stage of the food chain -- from carriers to equipment makers to
startups -- are violently overhauling their business strategy and
fighting simply to survive. What is emerging is an industry barely
recognizable from what it was just 12 months ago. And broadband still
won't be widely available anytime soon.
Before the bottom fell out, carriers such as WorldCom, AT&T, Global
Crossing, and Qwest Communications (which ran those famous "every book"
ads) were buying warehouses of equipment to build out their voice and
data networks. In 1999, data traffic surpassed voice traffic for the
first time, and the following year, U.S. carriers spent a record $117
billion on new equipment, according to Merrill Lynch. The carriers
needed faster switches, bigger routers, boxes that could convert
electrical signals into pulses of light, and dense wave division
multiplexers that could then multiply the number of light waves being
pushed down each fiber-optic strand. They needed equipment that would
increase the data-carrying capacity of their networks by orders of
magnitude.
"The opportunity looked quite significant," remembers Mike Unger, former
president of Nortel's optical networks business and currently a
consultant who sits on the boards of several startups. "The Internet was
doubling every hundred days. People were able to start up companies to
meet the growing demand for bandwidth. It was easy to extrapolate from
that growth and believe the opportunity would continue for a number of
years, so everybody continued to make investments."
A kind of arms race developed in response to this seemingly insatiable
hunger for bandwidth. Trying to expand their markets and keep growing,
equipment makers sold their gear not just to the major carriers, but
also to less financially stable competitive local exchange carriers
(CLECs), Internet service providers, and Web hosting companies. This
created intense competition, along with plummeting equipment prices and
a glut of new capacity. During the past decade, the network capacity of
the Internet has grown nearly 10 times as much as traffic itself,
according to a study by Internet pioneer Lawrence Roberts.
But here's the catch: Bandwidth is not spread out evenly throughout the
network. There's way too much capacity in the long-haul pipes -- with
only a fraction of those fiber lines even being lit -- and unmet demand
in cities and suburbs where the people who could use that bandwidth
actually live and work. In a nutshell, says Steve Georgis, CEO of
startup Network Photonics, "bandwidth is in the core network, but it
still has to be distributed to the end customer." And -- a bigger
challenge -- it has to be done cheaply. Merrill estimates expenditures
on networking gear will decline 11 percent to $104 billion this year and
then fall another 19 percent in 2002, to $84 billion.
Unfortunately, putting capacity in the long-haul lines was the easy
part. Cash-strapped carriers are now demanding that, in addition to
boosting bandwidth, new network gear must also lower their operating
costs. And since bandwidth is a commodity, the new equipment has to
allow them to offer more valuable services to their customers, like
bandwidth-on-demand. To do that, they will need to build more flexible
networks that can be easily switched on and off.
Innovative startups such as Georgis's Network Photonics are tackling
these problems. The company has developed an extremely clever optical
switch technology that replaces as many as 10 7-foot-tall racks of
equipment with a single switch the size of a VCR. More impressive, the
smaller switch would sell for less than $1 million, compared with at
least $10 million for older equipment. (The technology combines a
reflective grating with a micromirror device that is much simpler than
other systems in development today.) Similarly, Iolon, an optical
startup backed by Kleiner Perkins, is trying to sell "tunable" lasers
that reduce costs for telecom carriers. Older optical equipment uses
dozens of individual, fixed lasers, one for each wavelength of light.
But Iolon's technology would replace those with lasers that could be
tuned to many different wavelengths, thus requiring fewer lasers that
can provide bandwidth far more flexibly.
Both of these technologies represent steps in the right direction.
Unfortunately, having great technology is only half the battle in
today's market. "Carriers don't want to take the risk of buying from a
startup," Georgis admits. Because of that, he has had to completely
change his strategy. When Network Photonics was founded in 1999, the
plan was to make optical switches and sell them directly to carriers. By
July 2001, the company had raised $139 million in venture funding on the
strength of that plan. But now carriers want one-stop shopping -- hot
new technology that's incorporated into gear from established equipment
makers. It's akin to consumers buying a car with an awesome stereo
already inside it rather than buying one separately from Harmon Kardon
and installing it themselves.
Back in the days of the bandwidth rush, most startups were focused on
building entire systems, even if they were niche products, rather than
just key components to be integrated into other companies' boxes.
Companies like Nortel or Lucent could respond to the intense competition
they faced from the venture-funded pip-squeaks in one of two ways: They
could design their own competing technology or simply buy the
threatening startup whole hog. Today, telecom equipment makers find it
more difficult to pursue this make-or-buy strategy. They can't make new
boxes as easily, since they've slashed their workforces to the bone, and
they can't make acquisitions because their stocks are in the tank. So,
like it or not, the big equipment guys are going to have to rely more
and more on outside component suppliers for new innovations.
This represents a huge change in the industry's ecosystem. If you think
about it, right now the telecom equipment industry is where the computer
industry was 15 years ago: dominated by vertically integrated companies
offering products assembled with proprietary parts. For the industry to
survive and thrive, it will have to follow the computer industry. That
means a switch to standardized parts, where companies focus on narrow
horizontal slices of the overall end market. Component suppliers such as
Network Photonics and Iolon that make optical chips or tunable lasers,
for instance, could increasingly sell to more than one equipment maker,
thus creating de facto industrywide standards. If that happens, the
network gear on the market could become smarter, cheaper, and easier to
deploy. But until then, the much-ballyhooed broadband future will remain
forever on the horizon.
For more information and related links, see the online version of this
story at:
business2.com |
