I would like to know what are the drawbacks of using a same point source in the interface node to connect to various data sources via Rslinx OPC.
I believe that the interface will attempt to load all points with the same point source and than only use the points that match the interfaces ID. Basically, it means your interfaces will take longer to startup. Also, it might be more difficult to manage and track which interface ID corresponds to which OPC interface if you have multiple OPC interfaces running.
Dan Fishman is right, interfaces load by point source only and then filter by interface ID (location1) afterward. This can be really slow when loading thousands of tags over a slow network connection.
Also, it is easy for people to create a new tag with location1=1 when they really meant to use a different location1/interface ID.
To avoid confusion and improve efficiency, always use a different point source for every interface instance.
Back in the bad old early days of PI, the point source could only be one character, so you were limited to 26 letters, 10 digits and a few special characters like #. That's why T and C and A are still used in today's PI.
Does there still exist the issue where this happens anyway, for example an interface with point source OPCUK and one with OPCUSA would both have the same tags detected on startup because the PI API still only works with single letter point sources, or is my thinking out of date/never actually been true anyway (highly probable!)
I think you might be confusing PI API connection names and point sources. API application names by default are the first 4 characters plus 'E' unless otherwise configured to use the LONGPROCNAME in piclient.ini. I don't believe there was ever a truncation issue with point source name though there can be issues with application name truncation. The following KB's contain some more detailed information on API connection names.
I'm guessing you have been around PI for many years. On PI 2 servers and maybe the first PI 3 servers, point source could only be a single character!
I have seen some of this confusion first hand and recall around seven years, an OSIsoft instructor explaining (incorrectly) point sources should only be one character.
One last point to add to Dan Fishman's and Steve Boyko's answers is that some users like to search for PI Points using the Point Source as a search filter. Having a large number of PI Points linked to the same Point Source might make it hard for your users to locate the data they need. It makes more sense to have a unique Point Source per data source to easily keep track of where the data is coming from.
Hope this helps, Seb
Unique naming will certainly help - designate a numbering scheme based on the type of plant or region. OPC[1-10] for east side data sources, OPC[20-30] for west... That is how we do it. Interface ID could be using the same number scheme.
Retrieving data ...