Thanks Roger! As for a "solid alternative", I'm just curious for different approaches to the problem.
Specifically, it strikes me that Tor is limited by a much larger demand than supply for routers, and my initial instinct would be for an alternative protocol to make *all* nodes route -- a sort of "tit-for-tat" routing architecture. For example, lets say you and I both want to anonymize our traffic: with NAT penetration, we can help each other (I bounce my traffic off of you, and you do the same). If that could be achieved, then the number of routers scales with the number of clients. I'm sure there are all sorts of different tradeoffs that make it more or less suitable for different applications. -david On 07/21/2011 11:41 PM, Roger Dingledine wrote: > On Fri, Jul 15, 2011 at 05:37:53PM -0700, David Barrett wrote: >> So I'm using Tor for the first time in a project, but I'm finding I >> don't really understand it and can't find the right sort of >> documentation online. Can anybody help me understand: >> >> 1) How does Tor locate proxies? Is there some central database of Tor >> proxies that they register with on startup, and query at runtime? > > Yes: https://gitweb.torproject.org/torspec.git/blob/HEAD:/dir-spec.txt > > But see also > https://www.torproject.org/docs/bridges > >> 2) How do you force it to switch IPs at runtime? I've tried sending >> "signal NEWNYM" and that seems to return OK, but only rarely does it >> actually change the IP. The only semi-reliable method I've found is to >> restart the tor process. > > It depends what you're actually trying to do. If you're talking about web > browsing, your problem is likely because you're using polipo. The NEWNYM > signal tells Tor to use a new circuit for new stream requests. But since > polipo does aggressive keepalive, Tor never sees a new stream request -- > polipo just quietly reuses the same stream it had from last time. > > Firefox by itself also tends to do some sort of pipelining or > keepalive. You could switch to privoxy as your http proxy. It tends to > stick to http/1.0, which I guess is what you want here. > > The better answer is to integrate NEWNYM support with Torbutton so it > can instruct Firefox on what to do, and get polipo out of the picture > entirely. We're getting there. > >> 3) Are there solid Tor alternatives, especially any that incorporate NAT >> tunneling, or that provide a more programmatic interface (rather than a >> SOCKS/HTTP proxy)? > > For an interface for controlling Tor, you might like > https://gitweb.torproject.org/torspec.git/blob/HEAD:/control-spec.txt > along with things like > https://gitweb.torproject.org/pytorctl.git/tree > > So far we've resisted making Tor into a library, since there's quite a > bit of overhead in maintaining network state and circuits -- and some > of that overhead is borne by the network. I wouldn't want each user to > have twenty little mini Tor clients that didn't know about each other. > > As for "solid Tor alternatives", it depends what you want. Are you just > looking for something to implement an overlay network? Or do you actually > want good and well-understood anonymity properties? > > --Roger > > _______________________________________________ > p2p-hackers mailing list > p2p-hackers@lists.zooko.com > http://lists.zooko.com/mailman/listinfo/p2p-hackers _______________________________________________ p2p-hackers mailing list p2p-hackers@lists.zooko.com http://lists.zooko.com/mailman/listinfo/p2p-hackers
