A Network Manager’s job is to get data from Network Elements and present it to the administrators or operators. There are two ways of doing this activity: 1) Polling and 2) Trap. Here is a quick difference between the two:
Polling : A traditional way of providing operators with the network elements information. It’s characteristics are as follows:
- Pull Mechanism – Requests and get information from network elements at periodic intervals. The periodic interval is most often configurable.
- Provides non-real time information. It may happen that some changes happen in network element but polling happens an hour after that. Thus, operator gets to know about the changes after an hour.
- Higher bandwidth needed.
Traps : When an alarm situation exists a trap can be generated, or if some changes happen at network element, an attribute value change event can be generated by the agent. It’s characteristic are as follows:
- Push Mechanism – Event generated by network element in case of an alarm or attribute change
- Real time information
- Exceptions or alarms are reported
- Lower bandwidth needed
With the two mechanisms explained above, question remains which one is better. Preferred solution is Trap Directed Polling. When an alarm situation exists or some attribute value changes, a trap can be generated by the agent and on receipt of the traps, the manager can decide on what information to be poll. With trap directed polling the network bandwidth requirements are also very low.
However, direct polling can not be eliminated completely, some amount of polling is required to find the network performance.
Hope you find the information presented here useful and informative. Please leave your suggestions, feedback or query in the comments section below.