On 21 Nov (16:26:31), Tim Wilson-Brown - teor wrote:
On 21 Nov 2015, at 05:38, David Goulet dgoulet@ev0ke.net wrote:
I know that we have sometimes 1, 4, 5, 6 or 7 hops circuit but the general case is considered here and we have no stats on how frequent those unusual circuits are.
When I tested hidden service path lengths:
Most clients cannibalized a 3-hop path for their directory, introduction point, and rendezvous circuits. So 4-hop paths may be quite frequent on the client side of hidden services.
On the server side, it depends on how busy the hidden service is - whether it has any preemptively built paths to cannibalize or not. If so, it's side is typically 4 hops, if not, it is 3.
Indeed, I bet cannibalization happens more often then we think thus ending up with a 4 hops circuit to either your IP or RP.
It would be great to have some stats for typical path lengths, is there an open ticket for this, or should I create one?
That would help us have a better estimate of network capacity for sure but I wonder if it worth the efforts versus having a real privacy oriented statistics gathering system that could give us a much more accurate number of the used and unused capacity of relays.
In other words, question comes down to should we put effort in a bigger larger system or continue cherry-picking small stats here and there? (huge work once vs small/medium-ish effort multiple time :)
Thanks! David
Tim
Tim Wilson-Brown (teor)
teor2345 at gmail dot com PGP 968F094B
teor at blah dot im OTR CAD08081 9755866D 89E2A06F E3558B7F B5A9D14F
tor-dev mailing list tor-dev@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev