Couldn't find open source scripts to download and parse metrics from CollecTor server descriptor files so started creating some with help of ChatGPT. It'd be great if these already existed somewhere to build from and contribute back to! Created something very basic to parse and group relays by family fingerprint showing contact info, aggregated observed bandwidth, average observed bandwidth, unique IPv4 count, and unique AS count. Will keep dumping code here: https://github.com/1aeo/relayradar/
What would be helpful to see that you can't find elsewhere? Ex: I want to summarize total bandwidth observed by relay family per AS to see how different providers are performing
Source files: https://metrics.torproject.org/collector/recent/relay-descriptors/server-des...
Some initial insights, might be specifically incorrect but should have the right magnitude, from parsing March 27th, 28th and 29th server descriptor files: March 28th 6814 number of unique IPv4 addresses ~900 number of unique ASN Screenshot attached of example output in .html and attached .html file
Used a 3rd party API, ipinfo.io, to lookup AS per unique IP and they shared some more metrics below via images, based on the 3 days of server descriptor unique IP to AS lookup calls. Top 5 countries and cities, highly concentrated in Europe Top 5 companies, mostly hosting companies and 1AEO, us, as #3 (guessing by quantity of IP addresses) 38 mobile carrier IPs running relays
[image.png]
[image.png] [image.png] [image.png] [image.png]
tor-relays@lists.torproject.org