Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
 

Optimize the Network Management System Configuration for the Best Results

You can modify your network management system configuration to optimize the response time for SNMP queries. You can configure the network management system by following the below tips:

  • Change the Polling Method from Column-by-Column to Row-by-Row

    You can configure the network management system to use the row-by-row method for SNMP data polling. It is evident that row-by-row and multiple row-by-multiple-row polling methods are more efficient than column-by-column polling.

    By configuring the network management system to use the row-by-row data polling method, you can:

    • Poll the data for only one interface in a request instead of a single request polling data for multiple interfaces as in the case with column-by-column polling.

    • Reduces the risk of requests timing out.

  • Reduce the Number of Variable Bindings per PDU

    You can improve the response time for SNMP requests by reducing the number of variable bindings per protocol data unit (PDU). A request that polls for data related to multiple objects mapped to different index entries, translate into multiple requests at the device end. This is because the subagent might have to poll different modules to obtain data linked to different index entries.

    The recommended method is to ensure that a request has only objects linked to one index entry instead of multiple objects linked to different index entries.

    Note:

    If responses from a device are slow, avoid using the GetBulk option for the device, because a GetBulk request might contain objects that are linked to various index entries and might further increase the response time.

  • snmp bulk-get recommended number of OIDs and max-repetitions

    An SNMP bulk-get request responds with a total of (max-repetitions * number-of-OIDs) variable bindings. When interface statistics objects (such as ifInOctets, ifOutOctets, etc) are present in a query, the requests are sent to lower layers. Hence, there is an impact on the responses by an increase in the max-repetitions that you send in a bulk-get request. For bulk-get queries for interface stats objects, it is recommended to use the ‘max-repetitions’ value of 10, and the maximum number of OIDs per request is 10.

  • Increase Timeout Values in Polling and Discovery Intervals

    By increasing the timeout values for polling and discovery intervals, you can:

    • Increase the queuing time at the device end.

    • Reduce the number of throttle drops that occur because of the request timing out.

  • Reduce Incoming Packet Rate at the snmpd

    The following methods reduce the risk of SNMP requests piling up on any device.

    • Reduce the frequency of sending SNMP requests to a device.

    • Increase the polling interval.

    • Control the use of GetNext requests.

    • Reduce the number of polling stations per device.