Future IT: A Look at How It Will Evolve

Evolution is a hot topic in IT circles. There is, appropriately enough, evolutionary computation, which bases aspects of computing on biological systems that gradually change into “a different and usually more complex or better form.” The process of evolution provides models for dealing with the complexity of advanced IT systems. You could think of studying the development of species over time?or building models that replicate natural selection?as a giant Google search for the species (or IT application) that thrives rather than falters.

“Evolution itself is a fantastic search engine?it goes through millions and millions and millions of things, and comes up with extremely creative designs,” notes Eric Bonabeau, chairman and chief scientific officer at Icosystem, a company that does work in complex adaptive systems and counts DuPont, Humana, Intel and Schlumberger among its clients. Icosystem, for example, built a simulation that accounted for market conditions, employee scheduling and the rigors of humans working on oil and gas rigs for two straight weeks without much sleep to help a client determine the best way to manage its staff and equipment.

There are grander schemes to talk about, too. Future visions recently cited by academics such as W. Brian Arthur include a future where a smart traffic-signal system adapts instantaneously to minimize congestion and keep city motorists moving; or, in a poignant thought for millions affected by last summer’s blackout, an updated electrical grid that responds nimbly to surges in power demand during a heat wave.

Bonabeau says that the extreme complexity of today’s computing systems, and the often unseen connections between them, makes various aspects of our natural world potential models.

Two types of evolutionary approaches exist. One looks at evolution itself as the model, trying to map how things evolve. Hence “genetic algorithms,” which have offspring and mutate into more effective algorithms over time. The other looks at the products of evolution?biological systems. This holds that one should look to see how nature solves certain problems, such as communication in a bee hive or school of dolphins. The phenomenon stems in part from Janine Benyus’s 1997 book Biomimicry: Innovation Inspired by Nature, which showed how clever humans emulate nature to solve problems. For instance, there’s jigsaw computing, which models itself after neurons and other cells, and may provide a basis for developments in nanoscale computation. Photosynthesis and protein microtubules provide potential models for optical network architectures.

IT has long looked to the natural world for certain models. Expert systems and artificial intelligence were built to replicate how we think, and neural networks use the brain as a model. Various systems security efforts have drawn on the human immune system, notably IBM’s Digital Immune System for Cyberspace (currently licensed by Symantec) and Sana Security’s Primary Response product. And?who hasn’t heard of autonomic computers that will, among other things, heal themselves?

Perhaps it’s just coincidence that so many aspects of nature become popular in IT. After all, it’s only in the last few years that computing itself became second nature for most of us.