From...

January 28, 1999
Web posted at: 3:26 p.m. EST (2026 GMT)

by Jackie Cohen

(IDG) -- Chaos theory is something geeks talk about at parties to impress nongeeks. A well-placed copy of James Gleick's 1987 hit Chaos on your bookshelf also does the trick - regardless of whether you actually read the book, much less understand its topic.

Yet while Gleick proclaims that chaos can be found everywhere in nature and computing, few technologies other than artificial intelligence have really taken advantage of the powerful benefits of chaos theory.

Until recently, that is. Chaos theory, otherwise known as the study of fractals, is working its way into a handful of mainstream applications that show great promise for the Internet. Scientists at the Georgia Institute of Technology in Atlanta are developing communications and computing protocols based on chaos theory, described for the first time late last year in the journal Physical Review Letters.

MORE COMPUTING INTELLIGENCE
	IDG.net home page
Industry Standard home page
Industry Standard email newsletters
Industry Standard daily Media Grok
Industry Standard financial news
Industry Standard e-commerce stories
Fusion financial report Free registration required to access this site
InfoWorld Internet commerce section
Reviews & in-depth info at IDG.net
IDG.net's personal news page
Questions about computers? Let IDG.net's editors help you
Search IDG.net in 12 languages
Subscribe to IDG.net's free daily newsletter for computer industry cognoscenti
News Radio
	Fusion audio primers
	Computerworld Minute

Chaos theory emerged as a way to mathematically study the infinitely complex systems of the natural world. "Where chaos begins, classical science stops," says Gleick. The theory grew out of MIT computer science professor Ed Lorenz's discovery in the 1960s of equations with solutions that appear to be random.

"Chaos has come to be associated with the study of anything complex, but, in fact, the mathematical techniques are directly applicable only to simple systems that appear to be complex," writes Neil Gershenfeld in When Things Start to Think, published this month by Henry Holt.

One promising systems application of chaos theory is in the area of data transmission. Scientists at Georgia Tech claim to have sped up transmission using chaotic elements as the messengers. Rajarshi Roy, professor of physics at Georgia Tech, reports that he has sent data at 125 megabits per second - more than double what's possible with binary protocols - by using what he has labeled "chaos communications."

Roy uses a laser to generate chaotic wave forms - signals that are highly irregular, never repeat and don't use zeros and ones to transmit binary information. He then sends the information over optical fiber lines. Roy notes that chaos communications is much faster than binary means, and the use of irregular wave forms may prevent prying eyes from discerning the contents of a transmission.

"You don't know whether any information is being transmitted when you look at the [chaotic] wave form, so it could offer enhanced privacy," Roy says. From a commercial standpoint, this could provide a secure way to send credit card information over the Internet, notes Ken Carter, associate director of the Columbia Institute for Tele-Information, a research institute affiliated with Columbia University.

He adds that chaos communications could do wonders for bandwidth on the Internet. "It would eliminate any scarcity in the ability to transmit information," Carter says. "You would never have to wait for anything to download and you could use capacity on demand."

While bandwidth purveyors like Cisco say they have no plans today for using such applications of chaos theory, the future looks bright for it down the line. "Better bandwidth is obviously something that is appealing for lots of applications, particularly as the Web gets more and more traffic, and carries more types of transactions," notes Steve Bernstein, a manager at Cisco Systems in San Jose, Calif.

Georgia Tech's Roy admits that he has yet to learn of any side effects of such systems, such as whether fractal-generation guzzles system resources or whether the contents of a chaotic transmission are highly susceptible to such environmental factors as background noise.

To take his research to the next level, Roy is exploring a collaboration with France Telecom and John Pierre Godgebuer, a professor of physics at the University of Besancon in France, who are also pioneering research into chaos communications.

Meanwhile, one of Roy's colleagues at Georgia Tech, physics professor Bill Ditto, has even more ambitious, if far-fetched, plans: chaos computing. Ditto, who heads Georgia Tech's Applied Chaos Laboratory, has created a blueprint for building an operating system modeled on chaos theory, complete with a microprocessor chip that uses chaotic analysis to "evolve" its answers.

Ditto says the only requirements for building a chaotic operating system are two irregular or unpredictable elements. He hopes to use live human neurons, which he somehow would hook up to a computer.

While this plan might seem half-baked, there's significant upside. For one, it's a lot cheaper to cultivate neurons than it is to build silicon chips, says Ditto. And chaos-based computers might be better able to comprehend incomplete information. For example, you could speak commands in a casual, conversational way and your computer would understand.

"I do believe that chaos computing and chaos communications would be possible," says Galina Neumeier, an analyst at GM Market Link, a research firm in Atlanta. "Fractals are a very strong thing to apply to computing." But researchers still haven't figured out whether fractals are too potent for computing: No one knows what might happen if chaotic systems begin to think for themselves.