Reports of my death have been greatly exaggerated.
--Mark Twain, in a cable from London to the Associated Press
It's axiomatic that at any given time, someone somewhere on the Net is predicting
its imminent demise. If it's not the apocryphal modem tax (a rumor that recycles
around the Net about as often as Dave Barry's column on the beached whale they
blew up in Oregon or the long-since recovered Craig Shergold's dying request for
business cards), it's death by regulation, mismanagement, or lack of bandwidth, the
measure of how much data can flow across a connection at a time.
High bandwidth means lots of data; video and audio that a user can interact with in
real time, that is, instantaneously, are high-bandwidth applications. Barry's text
column is small, a page or so of text taking up only a few kilobytes of computer
memory or disk storage space; the three-minute video clip of the Oregon TV news
coverage of the exploding whale takes up 11.7 megabytes, enough memory or disk
space to hold twenty-three copies of this entire book, and demands a lot of
bandwidth if you want to be able to watch it as if your PC and the Internet were a
giant broadcasting system. When, in 1994, I retrieved the file via FTP, it took more
than an hour to transmit to my computer.
Bandwidth (and where more of it is going to come from) is intimately bound up with
the political and legal questions of equal access. The key issue is who is going to
pay for it. A growing lobby is claiming that the Internet's cooperative structure can't
work in a future of high-bandwidth applications such as real-time voice and video,
and that the Net will shortly be swamped by its own success if it doesn't "grow up"
and adopt a pricing structure that more accurately reflects the costs of expansion.
One of the more vocal members of this group is networking company 3Com
founder Bob Metcalfe, inventor of Ethernet, the world's most popular local area
networking standard, and now a columnist for Info World. In his acceptance speech
for the Electric Frontier Foundation's 1996 Pioneer award for his inventions,
Metcalfe hit the assembled Net lovers at the 1996 Computers, Freedom, and
Privacy Conference with a shocking prediction: the Internet would start to collapse
in 1996. Quizzed a little more closely, he modified this statement. It wouldn't exactly
collapse, but it would start experiencing the network equivalent of the blackouts and
brownouts the power grid sees when it gets overloaded. In the ensuing months,
when Netcom and America Online (AOL) both experienced major downtimes (AOL
for more than nineteen hours one day in August), a lot of people took that as
evidence that he was right.
But these networks are not the Internet; they are just some of the many networks
attached to the Internet. AOL later attributed its long outage to configuration errors:
the service was installing high-capacity switches within its local area network when
its network access provider, ANS, misconfigured routing information destined for
Because the entire system was down, the diagnostic systems that might have
caught the error were not in place, and it took many hours for technicians to identify
what was wrong and reconfigure the routers. Members were given a
free hour on the service in compensation, but as many commentators pointed out,
if you had come to depend on email, a free hour on the service doesn't go any
distance at all in recompensing you for a day's lost work.
However, it is probably accurate to say that AOL's problem might not have
happened in an earlier era, when its system was smaller and updates and tweaks
could be done by hand. It certainly would have mattered less with fewer people
dependent on the service specifically and email in general. In the automated
system now necessary because of the service's size, no one knew what had
happened. A couple of months earlier, Netcom was out for five hours with,
reportedly, a similar problem. These stories also highlighted how many points of
failure there are for Internet connections, which depend on the electricity supply
and telephone connections, as well as the reliability of the Internet service
providers (ISPs) on both ends of the correspondence.
In a paper presented at MIT in September 1996, research affiliate Sharon Eisner Gillett and Lotus founder and EFF co-founder Mitch Kapor examined the
Copyright © 1997-99 NYU Press. All rights reserved.
Reproduction in whole or in part in any form or medium without written permission of New York University Press is prohibited.
Be sure to visit the NYU Press Bookstore
[Design by NiceMedia]