Best practice approach? Use OPC Client objects compared to PCS

Hi,

PCS is our new approach to access data through the Communication Drivers and it is still evolving. Current limitations to be aware of is that IData currently is limited to approx. 50,000 data changes per second and we do not (yet) provide a built-in redundancy mechanism for it.

By comparison, using the OPC client to access the driver supports a higher throughput, and using the SuteLink client is even faster than that. The DI objects include the Redundant DI Object to provide redundancy and automatic failover between DI objects.

We are continuing to expand on PCS capabilities, to have the Communicaiton Drivers more tightly integrated to each product, for a simplified architecture, and to eliminate some of the extra configuration work currently required for DI objects. We're also expanding on communication patterns, such as the recently introduced IMethods interface which faciliates OPC UA Methods client and will soon also facilitate method calls to RTUs through AVEVA Telemetry Server.

Hej Richard,

Related question
If I have application server set up with redundancy and have OPC/UA set up in OI Gateway on each server, does PCS connection follow between machines when failover occur?

Regards
Ulrik

Hej Richard,

Related question
If I have application server set up with redundancy and have OPC/UA set up in OI Gateway on each server, does PCS connection follow between machines when failover occur?

Regards
Ulrik

No, one of the limitations is that we have not yet completed redundancy for PCS communication channels. Each OI Server instance must have a separate (and unique) "scope". I guess it is possible to implement your own switch between instances by updating the scope part of each IO reference upon failover, but my recommendation is to stick with SuiteLink (or OPC) for now.

AVEVA Telemetry Server achieves a redundant PCS provider by using the discovery mechanism, i.e. the second server gets discovered if the primary disappears, but for the OI Servers we intend to build redundancy directly into the PCS protocol stack. 

Thank you for the update Richard,

How does the roadmap for this implementation look?
Is it planned? or something in the future pile?

Best Regards

/Ulrik

It is under consideration, with relatively high priority, so I'd tag it as "near future" for now. I cannot commit to a release date or version yet.