On August
10, 2000, The New York Times authoritatively reported
that Internet traffic doubles every 100 days. It was
not alone in projecting Kudzu-like exponential growth
for the Internet. A similar statistic had been cited
by a host of gurus of the Internet, including George
Gilder, Jack Grubman and Tom Peters. Its credibility
had been greatly enhanced by the government. First,
in 1998,the Commerce Department, in its first major
study on its economic impact , reported, "Internet traffic
doubles every 100 days," then the Federal Communications
Commission echoed the same mantra.
If this claim was true, the amount of Internet traffic
would increase more than a thousand percent a year.
Such exponential growth, when extrapolated over the
next six years, would increase the number of responses
to hits from 100 million a day in 2000 to 100 trillion
in 2006. Even if one assumed 2 billion people would
have computers by then, and spend 24 hours a day on
them, they each would have to respond 30 times a minute
to provide such traffic, and them such proliferation,
doubling every 100 days, would leave little time for
non-Internet activities.
As improbable as such growth might seem, it became,
as The Economist put it (July 2002) "the industry's
favorite statistic" and " an essential ingredient of
dotcom business plans and conference slide-shows."
In anticipation of this projected traffic, telecommunication
companies, such as Global Crossing, laid hundreds of
thousands of miles of fiber optic cables both on land
and under oceans, and arbitragers, such as Enron, organized
markets to trade broadband capacity like any other scarce
commodity.
As seductive as the statistic that Internet traffic
mushrooms to twice its size every 100 days may have
been to Wall Street speculators, banks and venture capitalists,
it is a fictoid. The extent that the projected kudzu-like
traffic failed to materialize is reflected by two dismal
measures: the actual utilization of the fiber optic
cables constructed to carry Internet traffic, and the
actual price of bandwidth for this traffic. In 2002,
over 97 percent of these fiber optic networks— which
were pipes filled with silicon (or, in many cases, empty
holes)-- had no traffic on them whatsoever. The price
of bandwidth had fallen meanwhile by nearly 70 percent.
How did such a fictoid take root in the corporate
mind and its media accomplice? Andrew Odlyzko, a former
ATT researcher, who first exposed
it in November 2000, traced the origin of
the statistic to a subsidiary of WorldCom, called UUNET.
UUNET, which described itself as the "backbone of the
Internet," reported in the mid nineteen-nineties that
it had been doubling its own capacity (or over-capacity)
every few months to avoid a "capacity crunch." And that
part may have been true since UUNET counted in this
measure all the pipes for fiber optics— empty or full—
that WorldCom was building and swapping to increase
its capacity. Both the Commerce Department and Federal
Communications Commission based their statistic about
growth on the UUNET data about capacity. For a brief
period in the early nineties, while the Internet was
being transformed from a academic research resource
to a mass marketing media with cookies, spam and cheap
access, traffic may have increased at an exponential
rate. But as Odlyzko amply demonstrated, Internet traffic
could not have been doubling tri-annually when that
phenomena was reported by the Department of Commerce
in 1998, or when it was endlessly repeated by the Internet
Gurus and entered into the media's clip files unchallenged.
By this time, it was fully detached from any reality.
|