3.5.2 Monitoring ISDN Interfaces

Using general SNMP monitoring methods, there are two ways to monitor the ISDN interfaces for fault detection. The first way requires that the management station poll the managed node for the ISDN tables at regular intervals. When a change in status has been detected, the management station can take appropriate action to notify the user of a detected fault.

The second way is more efficient, but less reliable. This method requires the management station to act on received ISDN traps (see Section 3.5.1.3, "ISDN Traps"). Each time a B- or D-channel changes state, a trap is generated by the agent and sent to the management station. This trap contains the index and status of the channel that changed state. An important limitation of SNMP version 1 traps is that they are sent using an unreliable delivery method.

A third, more common approach that overcomes the unreliability of trap reception is to use trap-based polling. Trap-based polling requires the management station to poll the ISDN tables at regular intervals. This polling guarantees that a fault condition will be detected at each interval boundary. The reception of the ISDN traps optimizes the polling by forcing the end of an interval at the reception of a trap causing the fault to be detected immediately and the next polling interval to begin.

When providing fault detection capabilities on the management station, the most important table columns are the dlgIsdnSigLapdOperStatus and dlgIsdnBearerStatus columns. Because there are far more B-channels on a system than D-channels, it may only be practical to poll the dlgIsdnSigLapdOperStatus columns to detect faults. Inspecting the status of individual B-channels should only be performed when the B-channel is of particular interest, such as diagnosing a call connection.

Click here to contact Telecom Support Resources