Currently SharpRemote only measures average transport latency+service time (I.e ping), but doesn't offer more detailed measurements.
In order to track SharpRemote's performance over time more detailed measurements should be performed and stored, so they can be compared to each other.
Measurements shall include the latency induced by subsystem (serialization, jit, rpc) and be stored in HdrHistogram or similar in order to allow for more meaningful visualizations than just average or median.
A possible implementation might consist of a separate application "Profiler" which reads ETW events emitted by SharpRemote.
Another approach is to measure these latencies in SharpRemote directly and log them via dedicated loggers.
Currently SharpRemote only measures average transport latency+service time (I.e ping), but doesn't offer more detailed measurements.
In order to track SharpRemote's performance over time more detailed measurements should be performed and stored, so they can be compared to each other.
Measurements shall include the latency induced by subsystem (serialization, jit, rpc) and be stored in HdrHistogram or similar in order to allow for more meaningful visualizations than just average or median.
A possible implementation might consist of a separate application "Profiler" which reads ETW events emitted by SharpRemote.
Another approach is to measure these latencies in SharpRemote directly and log them via dedicated loggers.