[tor-talk] Increase tor HS reachability
arma at mit.edu
Mon Dec 5 21:11:34 UTC 2011
On Mon, Dec 05, 2011 at 09:31:27PM +0100, Fabio Pietrosanti (naif) wrote:
> sometimes it happen that some Tor HS it's difficult to be reached.
> I experienced a situation where from one Tor Client i've been able to
> reach a Tor HS, while from another Tor Client i was not able to connect.
> To achieve maximum Tor HS reachability, would be reasonable to have:
> - several independent Tor Client running
> - connect simultaneously to a Tor HS
> - the first that get a connection win (and it's supposed also to the
> fastest path, because it got the connect first)
> It's a sort of "Interface Bonding" technique, that try to automatically
> discover a faster path and increase network reliability.
> What do you think about it?
I think it will end up hurting more than it helps. One of the main reasons
it's hard to reach hidden services these days is that the introduction
points for popular hidden services are overwhelmed by circuit creation
requests from users who are trying to introduce themselves (that's step 4
If we changed it so every client that's trying to access the hidden
service mobs the introduction points even more, we will make the problem
We also have some bugs that make a single client put too much load on
the introduction point by trying over and over.
I think the fix is that we need to put *less* load on the introduction
points, and that will make them work better. Robert Ransom has been
working on this area lately.
More information about the tor-talk