White Papers
-
NerveCenter 3.7: An Overview - Overview of NerveCenter - How NerveCenter Correlates Events -
NerveCenter obtains data from SNMP agents running on managed nodes by processing incoming SNMP traps and polling the nodes for specific MIB values using SNMP Get requests. When a predefined network condition is detected, NerveCenter stores the event information in a finite state machine called an alarm. The alarm continues to track the status of the interface, node, or enterprise being monitored. The alarm waits for subsequent events or issues polls to determine if the condition warrants further action.
For example, by tracking events based on their persistence, NerveCenter can detect when a router experiences a persistent link-down condition. Upon detecting a downed communication link, NerveCenter waits to see if the link is restored within a given amount of time. If the link comes back up within the specified time, NerveCenter acknowledges the situation and returns the alarm to its normal (ground) state. However, if the link remains down, NerveCenter can perform a variety of actions based on your management strategy.
In another example, NerveCenter can detect when a group of nodes appears to be down or unreachable and then poll their parent router. If the router is down, further polling of those nodes is disabled until the router is back up.
To correlate and filter this data, NerveCenter relies on configurable models of network and system behavior, or behavior models, for each type of managed resource.
| Behavior model |
|---|
A behavior model is a group of NerveCenter objects that detect and handle a particular network or system behavior. A typical behavior model consists of an alarm with all its supporting polls and masks, though behavior models can have multiple alarms. Any managed device can be associated with one or more behavior models. For a description of behavior models, see How Behavior Models Work. |
Once a NerveCenter behavior model has identified a problem, it can take automatic actions, including notifying an administrator or a network management platform, executing a program or script, modifying the node's properties, changing SNMP values, and logging the critical data.
| Previous: What NerveCenter Does | Next: How Behavior Models Work |
| Please send comments or corrections to Information Development | This file was last updated on 10 October 2000 |