Internet latency and routing observations
We publish what we measure ourselves: round-trip latency from our own probes and the routing anomalies detected from those measurements. Free to download, free to redistribute, attribution required.
Why this data is different
Most public measurement datasets are thin outside Western Europe and North America. Our probes sit in places that are usually blank on those maps, so the observations here cover paths that other datasets simply do not sample.
Files
Daily latency observations
One row per day, probe city and target: sample count, median, 90th percentile, minimum round-trip time and average hop count.
Detected routing events
Latency spikes and routing anomalies our detectors flagged: type, severity, cities involved, current versus baseline latency, and the ratio that triggered detection.
Machine-readable manifest · Coverage: 2026-10-10 .. 2026-10-24 · Updated: 2026-10-24
License and attribution
Published under Creative Commons Attribution 4.0. Use it commercially, modify it, redistribute it. The only condition is credit with a link back.
How to cite
For a paper, cite the archived version on Zenodo: it has a permanent DOI, its files never change, and it holds the full history since June 2026. New versions are added under the same DOI, which always opens the latest one.
doi.org/10.5281/zenodo.22944461
For the files on this page, which are refreshed daily:
What is not here
Cable names, routes and ownership shown elsewhere on this site come from TeleGeography. That is their dataset, not ours, so we do not redistribute it here. Probe and intermediate hop IP addresses are also deliberately omitted.
Updates
Files are regenerated daily at the URLs above, so they stay valid in a citation or a script. The manifest carries row counts, checksums and the generation timestamp.