zoomkat:
...Perhaps starting with a simple http client/server setup might be simpler and quicker.
Perhaps. But with the Arduino Ethernet library, http is transported with tcp, which requires the application program to establish a connection, and is, therefore not "simpler and quicker" than UDP at the phy or link layer (I think).
Perhaps you will submit two complete example sketches that use a command at one end to turn on (and off) an LED at the other end and show whether it is "simpler and quicker" at the application layer level. I am really interested.
Of course a "real" application like this would have the sensor board verify the returned acknowledgment and repeat the command if it didn't "take" (or if it timed out). That would also be required (in my opinion) if TCP were used. I know that TCP guarantees delivery of the message (with automatic repeats if no acknowledgment is received), but that doesn't guarantee the message was received correctly. It still needs some extra logic in the application. So...I'm trying to visualize at what level it is "simpler and quicker" to implement a web server and web client instead of UDP "Output Board" and a UDP "Input Board."
I mean it's not a contest (I don't do contests), but the actual networking part of the sketches I whipped out for this thread is something like a dozen lines of actual code each out of about 60 or so lines of actual code (if you leave out all those comments and empty lines---I like to see about 2:1 ratio of complete program to uncommented-unwhite-spaced bedrock code, and that's what those have).
Regards,
Dave
Footnote:
As for the voices: It's a curse. Sometimes I just can't help myself.