Jemmac Software
OPCFailover logo. Deployment Scenarios

OPCFailover needs to be installed on a node where you have deployed an OPC Client application that would benefit from the OPCFailover capabilities.

The data source (maybe a DCS) must be accessible from at least two OPC Servers, both of which must be installed and operational on the same network as the OPC Client.

A Simple Failover Pair

Step 1
Verify that the OPC Client has access to OPC data when connected directly to each real OPC Server. Follow the OPC Server and OPC Client vendors instructions to setup this connectivity, in particular pay attention to the required security configurations. Perform appropriate read/write tests to the real OPC Servers.
Step 2
Use the OPCFailover Manager to create a failover group that contains all of these real OPC Servers.
Step 3
Re-configure the OPC Client to use the ProgID that references the failover group. When your OPC Client performs a connect it will connect to the currently active OPC Server listed within the failover group. Perform appropriate read/write tests to verify the OPC Client has full OPC data access.
Step 4
Stop the currently active OPC Server and verify that the OPC Client detects the communication problem and attempts to reconnect to the same failover group ProgID. The OPC Client should have full access to its OPC data from the alternate OPC Server. Again, perform appropriate read/write tests to verify the OPC Client has full OPC data access.
Step 5
Restart the OPC Server that you stopped and stop the active OPC Server. Verify that the OPC Client fails back to the originally active OPC Server and still has access to OPC data.
You have now setup and tested your OPC Client in an OPCFailover configuration!

More complex deployments

Really there are no complex scenarios. You might configure more than two OPC Servers in a single failover group, and you might configure more than one failover group.

Neither configuration can be considered as complex.

Note : You may well have OPC Clients on more than one node on your network. You will need to deploy OPCFailover to each of these nodes in a way similar to that described above.

Abstracting common ProgIDs

Some OPC Server vendors ship OPC Servers with common ProgIDs, e.g. 'DCSabc.OPCserver'. An OPC Client that wants to connect to more than one of these OPC Servers has to step up and override the RemoteServerName associated with the ProgID when it attempts to connect to that OPC Server. Many OPC Clients are not able (coded) to do this and so are not able to communicate with more than one such real OPC Server that is addressed via a common ProgID.

OPCFailover allows you to abstract the common ProgID to local, unique ProgID. These unique ProgIDs can be used by an OPC Client to communicate directly with a remote OPC Server normally rechable via its common ProgID. Since the new local ProgIDs are unique, the OPC Client is able to connect to more than one rela OPC Server at the same time. This does not need to involve any actual failover capability.

For example, setup two failover groups as follows:

Failover group 1 = 'DCSabc-a'
Member OPC Server 'DCSabc.OPCserver' located on Node-a1
Failover group 2 = 'DCSabc-b'
Member OPC Server 'DCSabc.OPCserver' located on Node-b1
The OPC Client can now use both ProgIDs 'DCSabc-a' and 'DCSabc-b' and not get involved with overriding the RemoteServerNode for any ProgID. Note, that in the above configuration failover is not configured; this configuration is simply enabling an OPC Client access to two OPC Servers normally accessible via the common ProgID 'DCSabc.OPCserver'.

You could extend the configuration above to enable failover as follows:

Failover group 1 = 'DCSabc-a'
Member OPC Server 'DCSabc.OPCserver' located on Node-a1
Member OPC Server 'DCSabc.OPCserver' located on Node-a2
Failover group 2 = 'DCSabc-b'
Member OPC Server 'DCSabc.OPCserver' located on Node-b1
Member OPC Server 'DCSabc.OPCserver' located on Node-b2