Hello
We have been testing the nRF SDK for mesh and we have implemented a roundtrip delay counter by sending a few messages with a payload of 8bytes eg. "HiServer"
Its a modified version of the light_switch example (simple on off)
The client fetches the time from log_timestamp_get and sends a message to a specified server, the server responds with a "1"
When the client receives the response it does another log_timestamp_get and we get the time from there ((time2-time1)/32768)
We then repeat this action for a number of times to get an average
The rountrip time with 0 relay node in between the two is 0,55s
The rountrip time with 1 relay node in between the two is 1,87s
The rountrip time with 2 relay node in between the two is 3,57s
The rountrip time with 3 relay node in between the two is 5,36s
These times do not seem reasonable, what could we be doing wrong? No changes have been made to intervals(its 20ms)
I've attached the code(main client, main server, models for both and model include files for both)
We have also tried looping a message from client to server to measure throughput but the program freezes after sending around 2 or 3 messages. How would one go about sending lets say 15-100 messages in a row to measure throughput?
I do not want to use the health messages as suggested in this thread .