On 14 Jan 2019, at 09:32, ronqtorrelays@xxxxxxxxxx wrote:
It's a unitless amount due to scaling, but it starts as kilobytes per second.
Your relays are on the south and west coasts of the US, which means they're further away from the Tor bandwidth authorities in northern North America and north western Europe. Tor load-balances for client latency and bandwidth capacity. Relays with higher latency or lower bandwidth are only partly used. But this reserve capacity helps during peak times. Tor also load-balances according to relay position in the circuit. Tor guards currently have about 200 Gbps capacity, and clients are currently using 75 Gbps, or 37%: So your guards have slightly lower than average utilisation: 1 Mbps / 5.1 Mbps = 20% 1.2 Mbps / 7.4 Mbps = 16% I wouldn't worry about it too much.
We're focused on migrating to a stable bandwidth measurement system for the next year or so. A failed bandwidth measurement system is even worse: then relays can just claim to be as big as they want to be. After that, we'll look at geographical dispersion. But if we spread the relay load out too far, client performance will suffer. (And users wont see reliable, consistent performance, which is even worse.)
Yes, Tor needs more exits, their utilisation is often close to 75%. T |
_______________________________________________ tor-relays mailing list tor-relays@xxxxxxxxxxxxxxxxxxxx https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays