On Sun, 16 Sep 2007 01:25:51 -0500 (CDT), Scott Bennett wrote: > I'd strongly recommend that you start with the tor overview > document at > > https://tor.eff.org/overview > > paying special attention to the cartoon describing how circuits are built, > which should begin to straighten you out on some of the other misconceptions > you've indicated regarding tor. To learn about the process in greater detail, > continue reading at > > http://tor.eff.org/svn/trunk/doc/spec/path-spec.txt > > To understand how tor clients (and servers) know what choices of servers > are available, you need to read the directory protocol document(s) appropriate > to the version of tor you run. For 0.1.2.1[67], read > > http://tor.eff.org/svn/trunk/doc/spec/dir-spec-v2.txt > > For 0.2.0.6-alpha, read the above and > > http://tor.eff.org/svn/trunk/doc/spec/dir-spec.txt
Thanks Scott, I understand now that Tor client downloads network-status documents with descriptors of available onion routers and then chooses the routers for building circuits from that list. I understand that tor client connects directly only to entry nodes, and never makes a direct connection with middle or exit nodes (unless they're later used as entry nodes for different circuits). I understand that I can use firewall to control the entry nodes used (the firewall would prevent connecting to bad IPs, certain countries, etc). But I still do NOT see how Tor connections to entry nodes can be controlled with Squid. It would make sense to use Protowall (with a blocklist from bluetack.co.uk) to prevent connections to bad IP ranges. That way entry nodes run by various "bad" organizations will not be used. But I'm still left with a problem of how to avoid nodes from certain countries. What especially bothers me is when ALL THREE NODES are chosen from the same bad country. I would really like to avoid that. I hope solution for Windows will come soon.
