README.md

November 20, 2023 ยท View on GitHub

This repository releases a dataset of 80 WiFi bandwidth traces collected in November 2023 in Beijing, China. Please feel free to use these traces.

Measurement Methodology

We used iPerf v3.1.3 to measure and record the WiFi link bandwidth. The client and server are located in Beijing and running Ubuntu 18.04. The server command is: iperf3 -s -p [SERVER_PORT] -D, and the client command is: iperf3 -c [SERVER_IP] -p [SERVER_PORT] -i 1 -d 200. The client runs Python scripts to perform each measurement (see scripts/).

Dataset Statistics

All traces (traces/) are collected in static indoor environments, including 4 locations: cafe, campus, office, and restaurant. Each trace lasts 200 seconds, with a granularity of 1 second. Detailed information is shown in the following table and figure (here, the bandwidth indicates the average value of each trace):

LocationWiFi InformationTrace CountBandwidth Range (Mbps)Avg. (Mbps)Std. (Mbps)
cafe5GHz WiFi 4 (802.11n)207.448~7.8677.8120.099
campus5GHz WiFi 5 (802.11ac)2036.709~73.16563.7509.754
office5GHz WiFi 5 (802.11ac)207.282~29.12415.5326.334
restaurant5GHz WiFi 6 (802.11ax)209.362~9.8079.5670.133

Dataset Format

File name format: wifi_[location]_[time].txt

  • [location]: ["cafe", "campus", "office", "restr"]
  • [time]: start time of a trace (GMT+8); the format in Python is "%y%m%d-%H%M%S"

Trace data format (each line in a trace): [Timestamp (s)]\t[Bandwidth (Mbps)]