In-Depth

Operations push threatens powerhouses

• By Jack Vaughan
• December 17, 2001

The company unveiled a version of its Topaz application performance management service that boasts a new diagnostic module for identifying probable causes of system performance problems. The Auto RCA module utilizes a statistical correlation engine (forged by Mercury in concert with Hebrew University) that applies baselines to identify abnormal data points which are analyzed in a system context according to predefined rules. The software is said to filter out irrelevant metrics, and to identify the likeliest root cause performance problems.

The software takes as its input Topaz Site Scope log and polling data, although it works with third-party monitors and frameworks as well. For now, company representatives are saying that the software is complementary to established systems monitoring solutions.

Topaz Auto RCA does not try to automate the whole job, said Yoav Banin, Topaz product manager. This echoes a common theme for Mercury product marketing, which seeks to provide broadly useful solutions rather than narrower offerings. Among 'narrower' offerings promoted by industry players in recent years are neural agent-based monitoring systems.

"We didn't feel [neural network] technology is ready, and we thought the development environments [for neural nets] are too complex," said Banin. Thus, Topaz RCA focuses on pre-configured rules that establish simple but meaningful relationships in operations data. The software tries to identify five to 10 common 'top suspects' in system slow downs, Banin said.

For more on E-Business, go to: http://www.adtmag.com/section.asp?section=specialreport.