SuitWatch/
Compaq Edition

Views on Linux in Business
by Doc Searls, Senior Editor of Linux Journal

Monday, December 10, 2001


Free Range Telecom

Linux is a free range operating system that runs on free range iron. Both are also most at home on the wide open range we call the Net.

These are facts that we need to keep pointing out, because as civilized creatures we remain — a century and a half since Walt Whitman called us on it — "demented with the mania of owning things."

Don't get me wrong. There is nothing wrong with Windows, Solaris or MacOS in the same way there is nothing wrong with a chicken raised in a cage or a veal raised in a pen. In fact, they are all, in their current versions, very highly domesticated operating systems in the most positive sense of the description. They just don't grow in the wild. Even Apple's new OS X, which was wisely built new from the bottom up on the wild & free form of Unix we call BSD, is about as natural as a board room upwards of its kernel, Darwin (which now roams the world as Yet Another BSD).

Civilization is at its best where it makes the most of nature. That's the challenge for hardware, software, and networking. The most productive forms of ownership in that context are inspired not just to control, but also to share: to constantly ask, "what's best for the market here?" Because markets are natural places too. Civilization itself grew naturally around markets.

Which brings me to telecom. Or to where telecom and markets meet. Because telecom lives in what it has politely calls "the regulatory environment." Real markets mostly operate outside that context. So far outside, in fact, that state governments in the U.S. are having a heck of a hard time imposing taxes on goods sold over the Internet by companies outside those state's borders. In fact, they're failing. This is good, because it helps us all understand how wild and free the Internet really is, and why that's a Good Thing. What we need now is for this lesson to sink into the FCC, especially on the matter of wireless.

Think of the Wired World as the shell out of which the great feathered bird of the truly wild and free Internet will eventually hatch. Nearly all the good stuff for which the Internet is so well known has happened over hard physical connections. Its physical infrastructure is still almost entirely copper and fiber. The beak through which the Net pecks out of its wired shell is WiFi (802.11). Unfortunately WiFi signals have less range than the average cordless phone. Yes, it's wonderful to connect wirelessly at your local Starbucks and in trade show halls. But the difference between WiFi and the wireless internet is the same as that between an aroma and a meal.

I see only two ways the Net can go wireless, and both are outside the regulatory envelope.

The first way is by datafying the cell phone system. This was the idea behind 3G, or 3rd Generation data telephony. But at its best 3G is still slow, and prospects for worldwide deployment are iffy at best. The better solutions involve improving signal quality where the infrastructure is already there. Arraycomm, for example, uses a technology that constantly maximizes signal strength exactly where you (or your laptop, or your handheld device) happens to be. This is done by applying computing intelligence to the combined signals of two or more existing antennas at each cell site. Arraycomm's CEO, Marty Cooper, is credited with having invented the mobile phone when he was back at Motorola. With "smart antenna" solutions like Arraycomm's, he says we'll get used to having multi-megabit bandwidth, whether we're sitting at coffee shop or riding in the back seat of a taxi. (Disclaimer: I've consulted Arraycomm.)

The second way is by freeing up spectrum.In A Cure for bandwidth blues,Kevin Werbach calls on the FCC to answer the need for Open Spectrum. Here's how he draws the big picture:

Through spread spectrum technology, many signals can share (or roam among) many frequencies using very low power signals, without crowding or interfering with each other. Kevin puts it this way:

He adds an ideal case for embedded Linux:

The problem is, we don't have the spectrum yet. WiFi's tiny slice of spectrum won't handle it. There is plenty of other spectrum available for reallocation, Kevin says. It wouldn't be hard to break loose enough to create a wide open range for the Net. If you want to read more about Open Spectrum, check out Kevin's Open Letter to the FCC on Spectrum Policy, and Open Spectrum: The Paradise of the Commons.

The wireless world is a naturally public space. The Internet doesn't need much room to break wild and free. Just enough for its native species todo what they do best.


Additional link: Compaq Telecom