GitLab

With the recent fuss about how Windows relays are easy to compromise, we should get some hard facts about what fraction of the Tor network runs Windows, and what fraction of the weights by capacity in the consensus would have you go through a Windows relay.

I expect it to be quite low -- Windows relays so far aren't often up long enough to attract a lot of attention, even if they do have the network capacity (which many don't).

Once we have that number, my follow-up question will be: what fraction of consensus weights for relays that have the Guard flag are running Windows? I expect it to be even lower, because of the barriers to getting the Guard flag.