OPC UA Redundancy Question

Ignition 8.1.10

I have a redundant Ignition gateway reads from and is read by a pair of redundant OPC servers. These servers are OPC DA and will use Matrikon Tunneller to communicate over our network. Diagram below.

I recently realized there are some OPC UA Redundancy configuration settings as of 8.1.10. I’m struggling to tell what the difference or benefit of enabling this redundancy is.

In my configuration, the master PI node would hand off to the backup PI node in the event the master gateway goes down. Would it just provide another failsafe such that both my tunnellers could read from the either Ignition gateway? Essentially creating this:

I have a redundant Ignition gateway reads from and is read by a pair of redundant OPC servers.

I don’t understand what you mean here.

Is Ignition the OPC UA client, is PI the OPC UA client, or are there connections going both ways?

Should have been more clear - there are connections both ways. PI (Matrikon) is acting as a client to read Ignition tags (no writing), and Ignition is acting as a client to read PI (Matrikon) tags. Read-only from both directions, effectively

Well, the settings you linked to only matter to 3rd party UA clients that understand what non-transparent redundancy is. It will create something like your 2nd diagram in that case, where if the Ignition master is down the client will be able to switch to the backup server, and know when to switch back.

I don’t know if the Matrikon tunneller supports this, but the OSI PI client does. Not sure how the tunneller plays into the client connections coming from PI.

Frankly I’m not sure whether tunneller supports it either, but you’ve given me enough of the necessary language that I can discuss it with folks on my end. Thanks!