Bandwidth Be Thy Name

The computer age is over."With these oracular words, George Gilder opens his new book, "Telecosm." It's in part a jeremiad against smug, slow-moving forces from the federal government to telephone companies to Microsoft, and in part a guide to what he sees as the next technological paradigm: a world of unlimited and nearly free bandwidth, created by a new Internet liberated by optical technology. This world will offer, he believes, unprecedented opportunity for both economic enrichment and personal freedom. "The computer era is falling," he writes, "before the one technological force that could surpass in impact the computer's ability to process and create information. That is communication, which is more essential to humanity than the computer is. Communication is the way we weave together a personality, a family, a business, a nation and a world."

You know George Gilder. He's the author of the 1981 book "Wealth and Poverty," a seminal document of the supply-side movement that gave us Reaganomics. His "Life After Television" (1994), which seemed to liken TV networks to the Kremlin, was by extension a prediction of the death of all old media. He is a peripatetic speaker at business retreats, and a contributor to Forbes. More preacher than pundit, he infuses everything he says with an almost religious zeal. "Futures are apprehended only in the prophetic mode of the inspired historian," he says in "Telecosm."

It's a curious book, oscillating from pages of dense prose about the physics of the electronic spectrum or "collision detection in a negative feedback loop" to poetic flights about the "crystal cathedrals" of light as a communications medium. Sometimes Gilder uses a word like "resonance" in a technical sense, sometimes in a metaphorical one, and he's not always clear about the difference. Long passages of exhortation give way to punctuations of journalistic anecdote. Still, he has an interesting argument to make, one worth pondering by anybody who hopes to transact anything at all over the Internet in the coming decade. And as always with George Gilder, he is often most useful precisely when he is at his least supportable.

The argument comes down to this: the Net as we know it is inadequate to the burdens we are placing on it. As one measure, a single Web site called filled 28 terabytes of storage in 45 days earlier this year--equal to the total monthly traffic of the Internet four years ago. Worse, more of the new traffic is "rich media": streaming audio and video, interactive advertising that draws on massive databases in real time, online gaming. But what about plans to increase bandwidth through such technologies as cable modems and digital subscriber lines (DSL)? Band-Aids, says Gilder; feeble retrofits that go unchallenged only because people are too polite, too ignorant or too heavily invested in amortizing existing physical plant. Take that, Bell Atlantic. Sayonara, AOL.

The network's problem is twofold. On the one hand is the predominantly server-client network architecture. It locates intelligence at the center of the network, treating PCs and other Web devices at the edge as dumb terminals. And it makes servers at the center a kind of ground zero for the bombardment of requests for data that now make up the Internet. On the other hand is an insuperable obstacle: the speed of light. When you call up a Web page, the response will be broken into "packets," each of which will be routed through a number of switching points, sometimes bouncing up to and down from satellites in fixed orbit 22,000 miles above the earth. Some packets arrive sooner than others, and in random order. They wait to be reassembled at your end. It all takes time. This time is unacceptable for rich media. For you to be happy, these transactions cannot afford a delay--or what engineers call "latency"--of more than 150 milliseconds, yet most Internet systems show average delays of 260 to 450 milliseconds. Over an obstructed Net there's a ripple effect, with milliseconds turning into seconds and even minutes, all due to the tyranny of light speed. As the old saying has it: "186,000 miles per second. Not just a good idea. It's the law."

Gilder's answer? An all-optical Internet. In this vision, the center of the network is dumb: unimaginable lengths of optical fiber and massive deployments of wireless optical hubs that serve as nothing but transmission media. Servers do not control the network. The intelligence is on the edge, with smart telephones and other devices that communicate with each other at speeds of up to 100 gigabits a second--almost 70,000 times faster than the top speed of a T-1 line. At these speeds, latency disappears. Communication anywhere on earth becomes by any perceptible standard simultaneous. Every single dream of interactive communication becomes possible. "Through these coruscating channels," he writes, "will ultimately run most of the commerce of the world; more value will move by resonant light than by all the world's supertankers, pipelines, eighteen-wheeler trucks and C-5A airships put together."


Is Gilder realistic? Technologically, why not? The carrying capacity of an optical network is everything Gilder says it is. Whether he is right about the proper path to an all- optical network (and he definitely has opinions), it can be achieved if both economic and political will exist. But each is in doubt. The private sector won't underwrite the optical network without some sense that markets will reward it; the current fate of cutting-edge companies is not encouraging. Governments, especially those in Europe, will have stomachaches about funding anything that seems likely to give disproportionate benefit to Americans. And, as always, there will be struggles over universal technical standards.

Then there's the question as to what consumers want from Gilder's telecosm. He's an unquenchable optimist. Regarding television as "a debauched cultural force," he assumes that if only we had the alternative of interactive TV, we would choose it in an instant. Disdainful of mass-market advertising, he believes that we would all welcome advertising tailored to our individual tastes. (He says nothing about the privacy problem, about how the advertisers get to know us well enough to know what we want.) The idea of the passive consumer--the couch potato--is anathema to him. He never asks whether consumers with access to the optical network would view it as nothing more than slightly improved television. He doesn't consider that advertisers would simply use the extra bandwidth to pour out even more manipulative spots.

In the end, Gilder is at war with himself. He upholds the dignity of ordinary people against big government and big business, and argues that technology will overcome both. But he seems unprepared for the possibility that people may not welcome their own liberation.

Bandwidth Be Thy Name | News