I guess polling is best in a Star topography

Thanks for the confirmation. Its clear now. Mine is a Star topography where I have one central node collecting data from many nodes around it.

Its possible that all the salve nodes try and send data to the Master when they have an update.

Its also possible that the master methodically requests for update from slaves one after other. If there is update they send or if not they send a dummy. Much better and can totally avoid collisions if otherwise two slaves try to send at once.

So I can send AT commands to the coordinator from the connected UNO to change the DL parameter to match the MY address of the desired slave - then send a request - wait for the response and then go to the next slave …repeat the process for all slaves in a circular fashion…

Request Slave01.
Get Response.
Delay
Request Slave02
Get Response.
Delay



Request SlaveNN
Delay
(Loop)

WIll the above be a good way to go ??

By all means yes, a node can send data to the central collector when you have data to be sent. All you need to do is provide it the data on its COM port and it will send it. Just note that if you are using sleeping modes, you need to use CTS hardware flow control otherwise you could loose your data.

Thanks… I do understand the point about sending data. But my query was on polling … where the co-ordinator requests many end points, one at a time, for data and gets it. This means that the co-ordinator will have to constantly keep changing the destination address in a pre-determined sequence. Guess it should be possible…

That would be correct.