You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on May 4, 2021. It is now read-only.
In one of the SBWS meeting chats the idea of measuring relays by pairing them with picks from a pool of fast relays was floated. I like the idea and have a suggestion:
For the fast measurement exit pool, require that relays also be guard-flagged. Guard+Exit relays are much faster and more reliable than Exit-Only relays. About 400 Guard+Exit relays are present with an average bandwidth of about 15MB/sec. An equal number of Exit-only relays exist with a much lower bandwidth average of 4MB/sec.
The text was updated successfully, but these errors were encountered:
I just want to note here that sbws currently picks a second hop like this.
It tries to find a 2nd hop that is 2x as fast, failing that 1.75x as fast, failing that 1.5x as fast, 1.25x as fast, 1x as fast, then finally failing that the fastest relay available.
The 2nd hop we are selecting is not always an exit: if the relay to be measured is an exit, we search for a non-exit.
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
In one of the SBWS meeting chats the idea of measuring relays by pairing them with picks from a pool of fast relays was floated. I like the idea and have a suggestion:
For the fast measurement exit pool, require that relays also be guard-flagged. Guard+Exit relays are much faster and more reliable than Exit-Only relays. About 400 Guard+Exit relays are present with an average bandwidth of about 15MB/sec. An equal number of Exit-only relays exist with a much lower bandwidth average of 4MB/sec.
The text was updated successfully, but these errors were encountered: