Tangentially related question:
Is there a big difference in terms of performance about reading half of the variables or, say, all of them?
I have a lot of machines with twin loading/unloading ports, but usually only one of them is working at any given time, so if it makes a big difference, I should split the nodes.
Also related to this:
If I have a DB with several variables (about 20-30) I need to write to implement an interface between each machine and some mobile robots that pick up the material carriers and deploy empty ones. In order to do so, I have to change a few booleans according to the interface steps, triggering them in order (that's how the supplier arranged it).
I guess according to your reply, either I create as many endpoints as variables are (a lot of work, having many machines) or I read the whole block from the PLC, modify only the values I need to modify, and then feed it back to the PLC.
Would that have a lot of impact, performance-wise?