Driver Cache Replication

The driver maintains a database of values received from devices known as the cache.

The cache is updated by both event and static data. When updating this cache with values received from the device, the active unit replicates these values to the non-active unit on a peer driver over the TCP link between the peer drivers.

To send the DNP3 acknowledge to the device after all received values have been replicated to the peer, use the Citect.ini parameters SyncReplicationDefault and SyncReplication.

This behavior is preferable (but not default) as it helps ensure that no data will be lost in case the active server dies in the instant straight after it acknowledges the transaction to the RTU but before the values are replicated to the peer. Also, the DNP3 acknowledge will be sent straight away if the peer TCP link is lost.

On cold start-up, if there is a running peer driver, the cache is synchronized from the running peer’s cache prior to the unit, on the starting driver, being allowed to become active. If no peer driver is found, or there is no redundant unit, the cache is cleared.

The cache value replication throughput is faster than 100 values per second provided that the underlying Windows operating system and TCP network does not introduce delays. This rate can be expected on a 10Mbit network that is not congested.

There is no disk file image of the driver cache. On cold start after a total system shutdown (shutdown of both peer drivers) the cache is empty.