RealTime IT News

IBM, Cisco Team on 'Self-Healing' Standards

IBM and Cisco broadened their partnership Friday by agreeing to work together on open "self-healing" software technologies and a corresponding standard to boost the performance of computer infrastructure.

Autonomic computing is the expression Armonk, N.Y.'s IBM has used to characterize computing infrastructure that regulates, or manages and "heals" itself in the face of network glitches or problems that arise from unexpected traffic spikes. This presents a time-crunch problem as some IT professionals said that one-quarter to a half of their IT resources are spent figuring out network problems, which typically span many disparate hardware and software pieces.

The latest advance in its autonomic computing efforts, which represent a key aspect of IBM's e-business on-demand strategy to grant customers greater control over the cost and availability of their foundation computing products, is a set of technologies and corresponding standards for crafting a common language to detect and system problems.

IBM and San Jose, Calif.'s Cisco will outline the initiative next week at IBM's annual CIO conference in San Diego, CA. Other vendors working on similar on-demand computing environments, such as Microsoft, HP, Sun Microsystems, CA and Veritas are all adding self-healing, or management features to their platforms to combat network issues.

IBM will integrate the new technologies into its broad portfolio of software, storage and enterprise hardware products right away in the first phase, said Dave Ogle, autonomic computing architect at IBM.

Cisco, who Ogle said was chosen to participate in this endeavor because of its leadership position as a top networking vendor, will jump in in Phase II, integrating the technologies into its products and services.

The new technologies deal with a another degree of autonomic computing functionality -- problem and fault determination technologies to "figure out exactly what the problem is and determine the next course of action," Ogle said.

"We aim to take a specific approach to fault determination where is a problems occurs, there ought to be common way of formatting data and defining the content of messages," Ogle said. "When a problem occurs in an environment, there is often a lot of finger-pointing to find out what or who went wrong. But with a standards-based approach, everyone is reporting same kind of things and you can start trying to correlate fixes in a more ad-hoc way."

Moreover, Ogle said IBM doesn't want this to be perceived as a proprietary technology set, so it has submitted a schema, Common Base Event (CBE), which the companies hope will serve as the basis for standardized, interoperable exchange of problem determination data via Web services, to the OASIS Web Services Data Management technical committee.

Ogle said problem determination is troublesome because of the multitude of ways that different parts of a system report events, conditions, errors, and alerts. Logging mechanisms contain a variety of content formats because systems are built using disparate pieces and parts with products from multiple vendors.

"The hygiene has to be done," Ogle said, explaining the need for an open standard. "so we don't 15 different ways to express the same thing."

IBM, who has also worked with Toshiba and Peregrine on autonomic implementations, has updated a number of technologies that are part of the mix, which may be examined on the company's AlphaWorks development site.

  • Log and Trace Analyzer for Autonomic Computing, a tool based on the Eclipse project that enables viewing and correlation of log files generated by IBM WebSphere Application Server, IBM HTTP Server, IBM DB2 Universal Database, and Apache HTTP Server. A practical example of a problem determination tool, the software makes it easier for developers to debug and resolve problems by converting disparate data into a common event model.
  • Generic Log Adapter for Autonomic Computing, a tool that transforms software log events into the situational event formats in the autonomic computing architecture. Consisting of a "rule builder" and configuration tool, and an adapter run-time environment, it provides a plug-in architecture for customization external to the user's software.

IBM's foray into open autonomic computing began in April with a blueprint based on open standards.