instead of waiting 20ms after every telegram sent over ip to fullfill 50p/s sending limit, the sent telegram in the last 1000ms are counted.
If the limit is reached, packets are dropped (this is unlikely).
The limit "rolls" every 100ms and is recalculated.
This allows sending "bursts" that exceed 5p/100ms but do not exceed 50p/s (this is within knx spec).
This could be further improved by a tx queue. But in most cases there is no need for that.
instead of waiting 20ms after every telegram sent over ip to fullfill 50p/s sending limit, the sent telegram in the last 1000ms are counted. If the limit is reached, packets are dropped (this is unlikely). The limit "rolls" every 100ms and is recalculated. This allows sending "bursts" that exceed 5p/100ms but do not exceed 50p/s (this is within knx spec).
This could be further improved by a tx queue. But in most cases there is no need for that.