NerveCenter^TM: Integration with Micromuse Netcool/OMNIbus - Contents - Overview - How NerveCenter Complements Netcool/OMNIbus -
What is NerveCenter?      Components Required for Integration

How NerveCenter Complements Netcool/OMNIbus

NerveCenter extends Netcool's ability to measure, predict, and improve service levels in several ways.

Smart Polling

Netcool/OMNIbus is designed to respond to events, such as SNMP traps, coming from managed resources. NerveCenter not only detects and filters SNMP traps, but it can poll resources at predefined intervals for specific network and system data. This information allows administrators to track network and system behavior and identify potential problems before they occur.

NerveCenter uses a feature called smart polling to minimize unnecessary network traffic. With smart polling, NerveCenter issues polls only when the outcome of the poll can trigger an alarm. For example, if a behavior model correlates high traffic followed by high error rates, a device is not polled for error rates unless it fulfills the high traffic condition. Using this same technology, NerveCenter is able to suppress polling to nodes that are unreachable because either they or their parent devices are down.

For service-level management, NerveCenter can help track the following statistical information over local and wide area networks:

• Who is on the network, what tasks they are performing, and what tools they are using
• What applications are in use at the application and device layers and how much bandwidth they are consuming
• What is the total volume of data on different parts of the network at the busiest time of the day and whether this traffic is seasonal
• How the traffic will grow with time, taking into account increases in the number of devices or users, and changes to the applications used

This information is essential to guarantee application and network service levels to both internal and external customers.

Intelligent Correlation

Event correlation is the mechanism by which NerveCenter evaluates a number of pre-defined events and determines how the events are related, what may have caused them, and whether the condition is serious enough to notify an administrator or take other corrective action. NerveCenter's complex correlation engine filters out redundant or mundane events so that only important messages are sent to probes. Reducing the number of messages sent to probes both facilitates network management and limits network traffic.

By reducing the amount of raw data sent to administrators, event correlation makes it easier for them to identify critical conditions quickly. Event correlation also results in the delivery of important information that administrators can use to establish baselines, monitor thresholds, determine network utilization patterns, track system performance, identify potential bottlenecks and other critical conditions, and plan for future network needs.

Distributed Architecture

NerveCenter's client-server architecture supports distributed polling across large networks. NerveCenter can be configured so that all polling is accomplished on local area networks rather than across a wide area network. Using this capability, you can reduce bandwidth and increase scalability by limiting the information to be monitored for each subnet and the number of nodes to be polled. NerveCenter servers running at remote sites can notify a centrally located NerveCenter Server or management platform of the noteworthy network conditions at those sites. Because the server can run as a daemon on UNIX systems or as a service on Windows, the branch NerveCenters can be managed remotely.

Other Advantages

NerveCenter offers the following additional advantages:

• NerveCenter has the ability to parse MIB values and obtain the variable bindings related to network events. You can extend the MIB values monitored as new devices are added to your network.
• NerveCenter's alarm actions include the ability to execute commands and scripts that can remedy the problem that caused the alarm, resulting in further reduction of the number of events that need to be reported to Netcool/OMNIbus.
• NerveCenter includes a set of predefined behavior models that you can use to monitor and manage your network. These behavior models contain all the required mask, poll, alarm, and property group definitions for basic network management using MIB-II objects. NerveCenter also ships with predefined vendor-specific models for monitoring Cisco, Compaq, and Wellfleet devices.
• NerveCenter includes a Web Client that makes it easy to monitor alarms from any machine that has a Web browser.
• NerveCenter includes tools for generating reports about the network.