Log In
Register
Forum
Topic RSS
9:32, EEST
June 25, 2014
Offline
Hello,
I’m new to OPC, so my question might be stupid, but I was wondering if there are any standards or best practices of how to set up Address Space (nodes) on server. I’m not going to set up server Address Space nor even manage the server myself, but I’m trying to find out the best way or provide the easiest way to define what nodes the client will read or be able to recommend a specific settings.
Let me give you an example. I have 3 machines with some common attributes (lets say temperature, pressure and state).
1) Should every attribute be stored in separate node?
2) Should the nodes be in specific hierarchy, for example:
– Machine 1
– Temperature
– Pressure
– State
– Machine 2
– Temperature
– Pressure
… etc
3) Should the name convension of nodes (node IDs) reflect its hierarchy?
4) Are there any performance advances in specific heirarchy / Address Space settings?
5) Are there any standards reflecting my questions? 
Thank you in anvance,
Best Regards,
Ladislav Asenbrener
13:15, EEST
December 21, 2011
Offline
1) Yes, every variable needs it’s own node, which can be a DataVariable or Property. They are both Variable nodes (NodeClass=Variable)
2) Yes, that looks good. A good convention is to define an ObjectType (for example YourMachineType) that defines a common structure. The Object instances (i.e. Machines) can then refer to the type using a HasTypeDefinition reference and tell the client applications what type of objects they are.
3) The NodeIds can be assigned freely. Often they are composed using the hierarchy, to ensure that they are unique. But you can also use for example random GUID identifiers (although they are harder to recognise and memorise for yourself).
4) No, I wouldn’t say so
5) The DI, ADI and FDI specifications, for example define standard device types. DI also defines some standards for the address space structuring: for example DeviceSet folder in which the devices should be located in the address space.
1 Guest(s)
