ISPPlanet Network Management System Series - NetPlus AutoManage

Device Control

SNMP is commonly used with read-only access to monitor traffic or error counters, but many devices also support read-write access to configurable parameters. In such cases, the NetPlus MIB Browser (NP_BROWZ) can be used to generate SNMP SET requests (right). Typically, you'd query the affected object(s) first. Next, select a result object and use "Change" to supply a desired value in ASCII or hex.

Finally, press "Set" to overwrite values in the agent's MIB. NP_BROWZ naturally rejects any attempt to modify objects that are specified as read-only, and stops when the first error is encountered in a multi-object update.

When considering SNMP for device control, it's important to realize that NP_BROWZ is just the wrench to turn the screw—if your device does not have an adjustment screw, it cannot be controlled through SNMP. Command line access through telnet is often used for device control, even when SNMP is used for device monitoring.

Network Monitoring
SNMP managers are commonly used to monitor networks in two fashions: by repeatedly polling devices with GET requests, and by listening for spontaneously-generated notifications known as TRAP or INFORM messages. The topology map created by NetPlus Node Discovery (NP_DISCO) provides both.

Before you begin monitoring, you'll probably want to configure a few parameters(left): the interval for polling all nodes, the color used for each problem severity (major/minor/critical), the severity assigned to each TRAP type, and associations between TRAPs.

NP_DISCO only reacts to incoming traps when it is also Monitoring (polling). When a poll times out or a trap is received, the source node and containing group(s) change color to reflect the severity of the situation. When a problem occurs, drill down from group to node, then right-click to view any TRAPs (right) from that node.Or use NP_BROWZ to GET or SET attributes to remedy the problem. If there is no associated TRAP, you must manually clear (acknowledge) or remove the TRAP before the affected node will revert to green (i.e., the color assigned to severity level none).

If you monitor a network of even modest size, you'll quickly come to appreciate the benefit of TRAP associations. Beware that this TRAP window is not automatically refreshed, and that you won't see any color changes unless you've mapped TRAPs to severity levels.

Network failures usually trigger a flurry of failed polls and TRAP messages involving several devices. To fix the problem, you must isolate the root cause. For this reason, it is handy to have a window where TRAPs from every device can be viewed together. This need is (partially) met by the NetPlus Trap Monitor (NP_TRAPS) application. One panel lets you receive incoming traps (right). Navigational buttons "scroll" back and forth between received TRAPs, but there is no window in which to simultaneously view several TRAPS or list all TRAPs.

Another NP_TRAPS panel lets you send traps to other SNMP managers (below, left). This feature can be used to manually forward received traps, although the source IP will be that of AutoManage. Other panels let you filter or automatically forward traps to other SNMP managers (below, right), based upon source, context, and prefix (the OID that identifies the TRAP). We found both filtering and forwarding a little clunky—for example, filters/forwards cannot be edited, new filters are not applied to existing TRAPs, and TRAPs cannot be searched.

go to part 4: Reporting / Final Words