IBM's autonomic goals

A number of vendors have strategies for bringing autonomic computing to the market, but each has its own vision for doing so.

IBM Corp. views the rollout of autonomic computing as a five-stage road map, in which Level 1 represents the basic state of having no coordinated systems management. Moving up, Level 2 provides centralized management tools but relies on humans to take action. In Level 3, systems can analyze problems and suggest changes. Level 4 offers the ability to correct problems without manual intervention. And Level 5 links autonomic computing to an organization's business objectives. Here, information technology management best practices are automated and become the policy basis for the autonomic system's actions.

IBM is now filling out that framework. "There are various capabilities available today that can get you to Level 4 and 5 in certain parts of the IT infrastructure," said Ric Telford, director of architecture and technology for Autonomic Computing at IBM.

Telford pointed to IBM Tivoli's Monitoring for Transaction Performance, which can allocate more servers and resources if performance degrades. The product works in conjunction with the Workload Manager functionality available on IBM's zSeries mainframe-class servers. Together, the products monitor transaction time and allocate additional resources if performance breaks down. Later this year, IBM plans to offer this self-optimizing feature on other servers — the pSeries and the iSeries, for example — according to company officials.