iperf3 Visualization Tool

This is a basic tool for visualizing one or more iperf3 log files using Vega-Lite and seaborn.

Example

A ribbon plot (line with error bands and scattered raw data) for an iperf3 log folder containing a week of logs (60 second log every 10 Minutes over a period of roughly a week).

Example (--box-plot)

Each box plot represents one iperf log file (60 seconds of measurement per file) measured every 10ish minutes over a time period of roughly 24 hours.

Usage

Data collection

Collect data using iperf3 and save it to any directory. Example: iperf3 -c server --json --logfile /path/to/data/file.json. If you experience high deviation during the measurement, run iperf3 longer using -t SECS (e.g. 20).

Requirements

Install the python requirements by pip install --upgrade -r pythonRequirements.txt. Do yourself a favor and use a virtual environment for that.

Visualize

Run run.py with arguments:

--data-directory /path/to/data/
--output /path/to/output/plot.png (optional)
--time-interval pandas_timeseries-offset-alias (optional, default H)
--box-plot (optional)

To generate the example ribbon plot:

python src/run.py --data-directory /path/to/speedtests --output ~/visualization.png --time-interval 1h

Note that this tool will skip all json files that contain the "error" key.

Contributing

Feel free to create issues or create pull requests. Clean-up, new features, bug fixes/reports welcome.

Thanks to @lassepe for implementing the Vega-Lite visualization in julia as a prototype.

rkost/iperf3Vis