IBM Corp has been trying to render human intervention in most tasks associated with systems management unnecessary. Describing its initiative as autonomic computing, the Big Blue has been working to create hardware and software with the "intelligence" to monitor and manage itself as part of a distributed computing system.

The goal of the exercise is to let systems resolve problems automatically.

The idea is to reduce the cost and complexity of operating computers by cutting back on the need for human administrators.

David Bartlett, vice-president of IBM Corp's autonomic computing division, said that to turn the vision into reality, through co-funding and indirect projects IBM had teamed up with labs in universities and partners in the industry.

It has created about 24 major co-funding projects while having over 1,000 projects developed by partners in several countries through indirect support. Its partners include Berkeley, the MIT, University of Texas, University of Michigan as well as units in China, Japan, Israel, India and elsewhere.

He said there were four areas of self-management IBM was concentrating on. The first is self-configuration, implying the systems can dynamically configure themselves to changing environments. This is followed by self-optimisation, which means components or systems continually seeking opportunities to improve their own performance and efficiency. This includes self-healing - the ability with which systems can automatically detect, diagnose and repair both software and hardware problems.

Finally, self-protection, which enables systems to defend themselves automatically against malicious attacks or cascading failures.

According to IBM, developing autonomic computing entails eight elements, among which are abilities to perform something akin to healing. Systems must be able to recover from routine and extraordinary events that might cause some of their parts to malfunction. They must be able to discover problems or potential problems, finding an alternative way of using resources or reconfiguring the system to keep functioning smoothly.

This includes the ability to detect, identify and protect against various attacks to maintain overall security and integrity.

Importantly, the autonomic computing system must know its environment and the context surrounding its activity, and act accordingly. It will find and generate rules for how best to interact with neighbouring systems. It will tap available resources, even negotiate their use by other systems of its under-utilised elements, changing both itself and its environment in the process - in a word, adapting.

Bartlett said that self-healing was the development of a single format. When the system fails it can decide whether the problem is in the application, operating system or database. The system must have the ability to pinpoint the problem before fixing it. The company aims at making it possible for businesses to have their computing systems working in accordance with their priorities.

A business, for example, can decide to have critical applications that mainly drive the revenue stream to be repaired first when the system fails. Following the pre-determined priorities, the system will then do all the self-configuration, self-healing and attend to other details itself.

Bartlett said that the immediate benefits would include reduced dependence on human intervention to maintain complex systems, accompanied by a substantial decrease in costs. Long-term benefits will allow individuals, organisations and businesses to collaborate on complex problem-solving.

Today, more than 475 product features in 75 distinct IBM products have autonomic computing capabilities. Each of these capabilities is based in part on self-configuring, healing, optimising or protecting technologies.

suchalee@nationgroup.com

Suchalee Pongprasert

The Nation

Melbourne