El mié, 23-07-2014 a las 18:34 -0400, Roger Dingledine escribió:

On Wed, Jul 23, 2014 at 11:24:47PM +0100, Noel David Torres Taño wrote:

...

What would happen if a Tor node changes behaviour and uses four or five relay steps instead of three?

Would it enhance Tor's security?

I assume you mean a Tor client?

https://www.torproject.org/docs/faq#ChoosePathLength

I read that and it gave me the idea. I'm asking what would happen to the network if somebody recompiles Tor to use e.g. 4 relay steps, and if it is really noticeable to somebody. As far as I understand it, the first hop does not know anything but that it must hand off a certain payload to a next relay, and that relay's address, but it does not know if the payload has two or three more onion layers.

...

Is it possible to relay Tor through a Tor connection? I mean using Tor with its three steps to reach a Tor entry node to get three extra steps.

Yes, it is possible. But it is currently considered a flaw, because it can be used to work around the 'infinite path length' defenses. http://freehaven.net/anonbib/#congestion-longpaths https://gitweb.torproject.org/torspec.git/blob/HEAD:/proposals/110-avoid-inf... https://trac.torproject.org/projects/tor/ticket/2667

...

Would that difficult correlation attacks?

Defending against correlation attacks is an open research, so "maybe". But it's not clear how it would, since an adversary who can see or measure your first hop (on the first circuit) and also your last hop (on the last circuit) would still be in the right place to do the attack.

I thought on that, but thought also that it may be more difficult to know which sites to monitor.

--Roger

Thanks Noel er Envite P.S. It seems it was silly, after all