| Description | The LMP notification rate depends on the size of the network,
the type of links, the network configuration, the
reliability of the network, etc.
When this MIB was designed, care was taken to minimize the
amount of notifications generated for LMP purposes. Wherever
possible, notifications are state driven, meaning that the
notifications are sent only when the system changes state.
The only notifications that are repeated and that could cause a
problem as far as congestion is concerned are the ones
associated with data link verification.
Without any considerations to handling of these
notifications, a problem may arise if the number of data
links is high. Since the data link verification notifications
can happen only once per data link per link verification
interval, the notification rate should be sustainable if one
chooses an appropriate link verification interval for a given
network configuration. For instance, a network of 100 nodes
with 5 links of 128 wavelengths each and a link verification
of 1 minute, where no more than 10% of the links failed at any
given time, would have 1 notification per second sent from
each node, or 100 notifications per second for the whole
network. The rest of the notifications are negligible
compared to this number.
To alleviate the congestion problem, the
lmpNotificationMaxRate object can be used to implement a
throttling mechanism. It is also possible to enable/disable
certain type of notifications.
This variable indicates the maximum number of
notifications issued per minute. If events occur
more rapidly, the implementation may simply fail to
emit these notifications during that period or may
queue them until an appropriate time. A value of 0
means that no throttling is applied and events may be
notified at the rate at which they occur.
Implementations should save the value of this object in
persistent memory so that it survives restarts or reboot. |