1 port multiple connections?

[mrbean]

That makes sense, it's just that I didn't see anything like that in the Multiple Connections - Server.vi.  Is that because that example assumes one PC for the server and clients?  What is the actual Spawning mechanism in LabVIEW?

[Mark_Yedinak]

Remember, the examples are not intended to be fully functional applications. They are basic code to illustrate a concept. The example as written would not work very well for lots of complex clients or clients sending lots of data. The example uses one tasks to process ALL of the client connections and that means that the connections will be serviced serially. In addition, a delay on one connection will impact the processing of other clients.

Anyway, take a look at the examples under "Programmatically Controlling VIs->Bynamically Loading and Calling VIs".

[mrbean]

Do I HAVE to have a server?  Can there just be two clients (where one is always listening)?  These are just very simple messages that need to go from the Test PC to the Unit Under Test over ethernet.

[Mark_Yedinak]

If you have a listener you by definition have a server. So, yes you MUST have a server. Whether or not you build that finctionality into one of your clients is up to you and how your application works. However, you will have a server (something that listens on a specified port for a connection).