# 06-02 Measurement Evaluation

Higher intervals between uploading measurements allots more CPU cycles to collecting measurements. In combination with higher backend timeouts to combat network latency shows that my assumptions yesterday were wrong.

I wrote a shell script to sort measurements by time, now the data collection process is fully automated. The server log file can be copied into a folder and a script gets measurements by origin (client, gateway, correction algorithm) and sorts the streams by timestamp. Here is a comparison of measurements at different distances and different gateway pool sizes. Higher pool sizes mean higher accuracy when calculating averages but increase latency and increase system load.

# 500 Measurement Average

Measurement stream at 1m with 500 measurement average reference value Measurement stream at 2m with 500 measurement average reference value Measurement stream at 4m with 500 measurement average reference value Measurement stream at 6m with 500 measurement average reference value

# 1000 Measurement Average

Measurement stream at 1m with 1000 measurement average reference value

← 06-03 Compile report 06-01 Timestamping results →