This defines how often a poll packet is sent to the device.

If once-only(1) is chosen, then when the RowStatus is set to active, a single poll packet is to be sent. There are no re-transmissions and no traps are generated. The results of the operation can be read in the pollLastPollSent and pollRoundTripTime fields. When a reply to the poll is received, then the RowStatus is changed to notInService.

For the other values, setting the RowStatus to active causes the device to be regularly polled. The operation of the poll can be monitored by reading the pollLastPollSent and pollRoundTripTime fields. After four un-successful polls a pollLostContact trap is generated. The next successful poll to the device causes a pollEstablishedContact trap to be generated. Changing RowStatus from the active state causes polling to stop - in particular changing RowStatus from active to notInService (and vice-versa) is a convenient way of stopping and starting the polls.

A poll is said to have been un-successful, if no reply has been received by the time the device is due to be polled again - i.e. if the pollRoundTripTime is still zero.

By convention, row 1 of the PollTable is only ever used for once-only(1) polls.

The facilities of the RMON Alarm and RMON Event groups are used when regular polling is taking place. These groups perform the timing, threshold checking and trap generation.