WASHINGTON, D.C. — IBM’s Tivoli division says it has harnessed a technology that will cut some of the grunt work out of IT maintenance, but it all started with a game of chess.
In 1997, IBM supercomputer Deep Blue beat reigning
world chess champ Garry Kasparov, sparking debate as to whether humans had finally built a computer that is smarter than we are.
At the time, George Plimpton, editor of The Paris Review, commented on IBM’s Web site: “”Over the years, it may be that the machine is likely to be able to do other things. I’m not sure what, at the moment. It can’t manage a baseball team.””
What IBM has been able to do with it is translate the intelligence into some of its products. “”It’s like a science project,”” says vice-president of products for IBM’s Tivoli division, Carl Kessler. “”Like any science project, it only becomes interesting when you can do something with it.””
The most recent iteration of that intelligence is called autonomic computing — named after the autonomic nervous system that regulates human functions like breathing. IBM uses that principle with some its hardware products under the eLiza code name. Tivoli has built the intelligence into 18 of its software releases, including Monitoring 5.1 and Access Manager 3.9 which were introduced at the Planet Tivoli user conference this week.
Autonomic computing promises to take human intervention out of some of the low-level problems of software and hardware management. “”Let’s say we’re in a Web application server and it’s running the IBM Tivoli Monitoring product,”” explains Kessler. “”If it detects a problem at the software level on the server . . . it can take automatic steps to repair it without having to say, ‘Can you send a human out to fix it?'””
Autonomic computing also refers to predictive computing: fixing problems before they actually happen. In the case of Monitoring 5.1, it can be used with Tivoli Enterprise Data Warehouse (released in April) to store all software incidents. This data can be analyzed to establish patterns and recognize common problems. Rigorous data collection is necessary in order to spot these trends and allow the software to step in before something goes wrong, said Bob Yellin, Tivoli’s vice-president of technology. “”Unless we work with a new paradigm, we are all going to be inundated with problems we can’t solve.””
There are IT managers who may blanch at the idea of software and hardware taking corrective steps without a little human prodding. “”We’re going to get that all the time,”” said Chris O’Connor, Tivoli’s director of performance and availability. “”If someone wants to go down the traditional path and build it from the ground up . . . they can go do that. (But) I tend to believe that even the most stubborn IT manager is going to end up cheating and looking at the (manual) and saying, ‘So how do they do it?'””
IT managers may appreciate the opportunity to let the software handled the drudge work, said Michael Roy, president of Vancouver-based Tivoli reseller Blue World Information Technology Inc. “”I think the idea is, you don’t get into cycles of dealing with lots of isolated support calls. I think there’s a market for (the software). If you talk to any big group that is dealing with complicated middleware and they always have some issues with support.””
Increasingly, IT departments are being asked to focus on broader business goals, he added, and spearhead the large projects, such as Web-enabling an enterprise and ensuring security is sound.
“”There is a shortage of skill needed to run really complex, real environments,”” he said. There’s lots of guys out there who are Microsoft certified and so on . . . but they aren’t necessarily looking at the key areas that require care and certainly require some action and planning to consolidate effort across platforms.””
The increasing complexity of IT means that it can never be fully automated, observed Arvind Krishna, Tivoli’s vice-president of security products. He sees current levels of autonomic computing dealing with the bottom 10 per cent of software maintenance. That could rise to 30 or 40 per cent in years to come, but there will always be more higher level tasks added to the top.
Besides, Kasparov only lost by one game.