The MQTT broker you install will likely have the capability to publish via a command line program. It's no worse than NC.I doubt using the MQTT functions will make this better (it probably makes the memory use even tighter), and it makes testing more complex because currently I can simply test by using the "nc" command on my Mac, but for MQTT I'll have to write python code on the Mac first.
One advantage of MQTT is that the server doesn't need a listing of listeners. If you decide to add a fifth listener, it won't require you to update anything on the server.
I also doubt the "overhead" of MQTT will be significant for this use case. It reads from a TCP socket and does some very light processing on the first byte or two of data in each message. It is a protocol that is designed for purposes exactly like the one at hand. The biggest problem with MQTT is reliability when you actually NEED messages to be delivered, and that doesn't apply here.
Statistics: Posted by jags84 — Mon Jul 22, 2024 12:31 pm