[tor-talk] tor-blocking sites

[Roger Dingledine]

On Mon, Feb 06, 2012 at 07:24:31PM +0000, Mr Dash Four wrote:
> Initially, there was a small number of these in the wild, but now it
> is widely spread - google is the main offender, but youtube (which
> is, as we all know, google-owned) and now, wait for it, scroogle.org
> (a site I use a lot) is also at it!

Google doesn't specifically single out exit relays. Google pops up a
captcha when a given IP address has asked it too many questions recently.
It's a defense mechanism against crawlers from, say, Bing who are trying
to steal their precious secret sauce.
https://www.torproject.org/docs/faq#GoogleCaptcha

> Tor-blocking could be very easily to implement by parsing
> cached-descriptors{.new} to see all exit nodes and then add them to
> a blacklist and start blocking. Is there anything which can be done
> to prevent this?
> 
> I am thinking of something similar to what is currently in existence
> with the bridge system - you don't know them all, just a portion of
> it, enough to connect you to the network. Could something similar be
> implemented with tor?

You may like https://www.torproject.org/docs/faq#HideExits

In theory the Tor protocol supports an "exit bridge" concept: the Tor
design separates the relaying component (how to establish a circuit given
a set of relays) from the discovery component (how to learn what relays
to use), so you could just bolt on another discovery mechanism. But
nobody has designed or implemented one.

--Roger