“With autonomic cloud computing, management of processing resources would be handled without human intervention.”

IBM, others team up on 'autonomic' computing

Computers that automatically heal and manage themselves might be reminiscent of the villainous "Hal" that turns on its human programmers in "2001: A Space Odyssey."

But researchers say so-called "autonomic" computers will be key in the future as businesses and other organizations increasingly rely on the Internet to manage their data and operations.

Georgia Tech University and Ohio State University plan to announce today a partnership with IBM Corp. researchers in Austin and Raleigh, N.C., to develop technologies to advance autonomic computing on the Internet.

The researchers will focus on "cloud" computing initiatives that combine the resources of computers in different locations using the Internet.

With cloud computing, a company or organization can use the Internet to split its computing workload between data centers in different parts of the world, or even outsource some of it to third-party companies.

Google Inc., Hewlett-Packard Co., Amazon.com Inc. and other companies all have launched major cloud computing initiatives.

At the same time, Georgia, Ohio and Texas have all emerged as hot spots for back-office data centers.

One of the problems with cloud computing, though, is managing thousands of computers and making sure they work how and when they're supposed to.

"The scale becomes impossible for a human" to track, said Georgia Tech's Karsten Schwan, director of the Center for Experimental Research in Computer Systems.

That's where autonomic computing comes in.

Named after the body's central nervous system, autonomic computing systems can react and make decisions on their own.

That's not as sinister as it may sound, researchers say.

One goal is to figure out ways to let computer data centers automatically call in more resources from other data centers when there's a spike in Web traffic, for instance.

Another goal is to design computers that can automatically shift work elsewhere to increase efficiency or save energy.

Finding ways to make computing more efficient is IBM's primary goal in the partnership, said Matt Ellis, vice president of autonomic computing at IBM.

Often, Ellis said, a company's computer servers might run at 20 percent or 30 percent of their capacity, even though the company is paying to power and cool them all the time.

With autonomic cloud computing, management of those processing resources would be handled without human intervention.

With such a system, a company could save money on new computers, power, cooling and technical personnel.

Officials from IBM's Austin TX based Tivoli software division are scheduled to be at Georgia Tech in Atlanta today for a ceremony formalizing the partnership.

Check Broadband-T1 Pricing Now >