NetCool Scalability Expanded

MicroMuse Inc. this week will address the scalability issues that have dogged large users of its NetCool/Omnibus event management software.

The network and systems management tool, widely used among service providers and major enterprises to manage IP networks, was redesigned to handle hundreds of concurrent operators and manage hundreds of millions of network alarms in a day.

Also driving the improved performance and scale of NetCool/Omnibus 3.5 are customers looking to keep up with an increasing number of security alarms and wireless service providers looking to manage their services as they scale, according to Brent Compton, director of product management at MicroMuse, in London. "Product scalability is a chronic issue," said Paul Bugala, an analyst at International Data Corp., in Framingham, Mass. "It gets more difficult as you try to incorporate more analysis in terms of root cause and add more elements to be managed."

For large-scale NetCool customers that use the tool for a variety of users, the performance boost is welcome. "We have project managers, command-center users and second-line support using our system," said a NetCool user, who asked not to be identified. That NetCool implementation at a large telephone company supports casual users working on hundreds of projects who access the NetCool system via Web browsers. "We use it for service-level views as well," said the user.

NetCool 3.5 now supports two-tier or three-tier hierarchical processing cores. The previous version was made up of a single object server and multiple desktops.

The new release allows multiple collection object servers to feed pre-processed data into the central processing core, which is accessed by desktops. The three-tier architecture includes collection object servers and the central object server and adds desktop servers that sit between the central object server and the desktops and distribute the read-only processing load.

MicroMuse is adding a NetCool option intended to speed mean time to repair. Virtual Operator automates responses to common problems, freeing personnel to work on harder problems. Virtual Operator is used for each pre-processing object server implemented in a NetCool system to cut the number of alarms that must be processed. That helps improve performance. At the same time, Virtual Operator can improve mean time to repair because it can work on multiple alarms simultaneously.

Virtual Operator, available now on Solaris, costs $35,000. A later release will support Linux, IBM AIX and HP-UX.