Unlike root cause services from system management companies such as Tivoli and BMC, the Mercury offering diagnoses problems in the user experience of the site, rather than problems apparent in the server, according to European product marketing manager Andy Crosby. "Other products measure the speed of response, but not whether the page returned is the right one, or whether other requests were lost," he said. "You shouldn't start diagnostics from the server."
Competing analysis products have a major claim to being better than TopazRCA, because they automate the whole process, diagnosing the problem and suggesting fixes. Topaz AutoRCA merely suggests a top five list of possible causes, but Crosby points out that this could be a better service, since the Mercury product includes monitoring of actual user transactions, and diagnostics are not reliable yet.
It is likely to sell first to existing Topaz users who are currently spending hours on analysis of problems, or whose applications lose large amounts of money per minute of downtime, said Crosby.
"During testing, one major UK building society used AutoRCA to find a problem, in a few minutes, that took the incumbent enterprise system 14 hours to find," said Crosby. "The system -- an inter-bank business-to-business application -- was losing Ã,£100,000 (about AU$300,000) an hour, so the company saved Ã,£1.4m (roughly AU$4.2m)."
Another benefit for existing Topaz users is the fact that AutoRCA works on historic data as well as live data, according to Crosby. Where the product picks up recurring problems with a service provider, some users may find it pays for itself in the money it allows them to claw back through service level agreements.
Mercury started life in 1989 with products that test enterprise applications and claims to have a 65 percent market share for its Load Runner load testing tool; after the Y2K panic raised the profile of testing to the board room, it moved to monitoring the performance of Web services.
