IBM tackles network troubleshooting

IBM is preparing software that promises to help companies more quickly troubleshoot glitches in networked computers, as part of its autonomic computing initiative.

Big Blue has released a beta, or test, developer tool for the "log and trace" tracking software. The software, which collects and collates historical data from multiple servers, is a product of IBM's autonomic computing effort to create systems that can manage themselves.

A critical component of that effort--and an ongoing aggravation for customers--is the ability to diagnose and fix problems that occur on servers, said Ric Telford, an autonomic computing director at IBM. System administrators spend a great deal of time simply deciphering and relating information from various log files in order to locate the source of a problem.

"It's a very time-consuming, human-intensive task. The value of autonomic computing is (in) making the system more self-managing and reducing complexity," Telford said.

By automating manual and repetitive tasks through autonomic computing, IBM hopes to help companies streamline their operations and free up information technology professionals for more valuable work. Rival projects from Sun Microsystems, with its N1 plan, and Hewlett-Packard, with its Utility Data Center, are also intended to increase use of customers' existing computers.

Typically, when a networked computing system fails, system administrators look at the log files of different components to determine how the glitch occurred. These logs automatically record the activities of both computers and software such as a Web server.

For example, if a Web site fails to complete a banking transaction or to deliver a Web page, IT professionals will look at the sequence of steps in the transaction to pinpoint the problem. The infrastructure involved may contain several components, such as a Web server, a database, and server software for running the application.

IBM's log and trace software is designed to put the log data from different software servers into a common format, so that administrators can track the sequence of events more quickly. Initially, the software will create a common format for log files from Apache Software Foundation Web servers, IBM's Web server and IBM's application server WebSphere. The company plans to extend that log file format to other appropriate products.

The software will be available first to application developers. Later this year, IBM's WebSphere Studio application development program will include the self-healing capability, which includes the problem resolution software, in the next two weeks.