Did some additional testing with a larger delay (20000ms) to take a 'control' look at message delivery latency for T-Mo -> T-Mo, and noticed that the send timestamps were not properly distanced. Many were, but there were outliers that will adversely affect analysis. Here is the difference in sent timestamps for 11 test messages:
msg01: n/a
msg02: 20035
msg03: 60476 <-- much too high
msg04: 26858
msg05: 20025
msg06: 124787 <-- whoa, nelly!
msg07: 20025
msg08: 34422 <-- too high
msg09: 22692
msg10: 20033
msg11: 20039
This sample shows a standard deviation of over 33000.
Did some additional testing with a larger delay (20000ms) to take a 'control' look at message delivery latency for T-Mo -> T-Mo, and noticed that the send timestamps were not properly distanced. Many were, but there were outliers that will adversely affect analysis. Here is the difference in sent timestamps for 11 test messages:
msg01: n/a
msg02: 20035
msg03: 60476 <-- much too high
msg04: 26858
msg05: 20025
msg06: 124787 <-- whoa, nelly!
msg07: 20025
msg08: 34422 <-- too high
msg09: 22692
msg10: 20033
msg11: 20039
This sample shows a standard deviation of over 33000.