Datacentre-latency-map

This is a first prototype of a datacentre latency map using RIPE Atlas.

It consists of:

  • A script to start measurements: start_measurements.py
  • A script to pull measurement data from RIPE Atlas and visualise it: process.py

Workflow

To start using this, one would need:

  • A RIPE Atlas account with enough credits to run measurements, and an api-key for that, stored in ~/.atlas/auth
  • A configuration file for the measurements conf.yaml , an example configuration is available in conf.yaml.example.
  • After running start_measurements.py, a summary of measurement metadata is written to msm.json in the local directory.
  • Once measurements are actually done (allow some 10-15 minutes for that), one can process the measurement data with process.py. process.py needs a local all-probes.geoloc.txt file (example format in all-probes.geoloc.txt.example), with RIPE Atlas probe IDs, with lat,lon,v4ASN,v6ASN and country information. You can run the getallprobes.py command to query the RIPE Atlas probe API to generate a recent file like this. process.py will create a ./map/geodata.js javascript file which has the relevant measurement data in a geojson-compatible format.
  • Point a browser at ./map/map.html to see the result.