[tor-talk] End-to-end correlation for fun and profit

[Maxim Kammerer]

On Tue, Aug 21, 2012 at 10:33 PM, The Doctor <drwho at virtadpt.net> wrote:
> That suggests that an attacker won't get nearly as much information with this attack as it seems.

No, you didn't understand the point of the attack.

> A valid point, I will concede.

Nothing written about /16 and /24 networks in this thread is relevant
to the (obvious) attack. I don't understand the purpose of most
replies here. How can it be described any simpler? You intercept
traffic at 25 nodes, you can deanonymize 25% of the traffic, or watch
80% of exit traffic. This makes one wonder whether this mailing list
is suitable for any discussion that assumes even the most basic
background.

> Exactly how strongly weighted by bandwidth is node selection these days?

It is linear, unless I missed something in the code. If it wouldn't be
linear, then nodes would be under- or over-utilized.

-- 
Maxim Kammerer
Liberté Linux: http://dee.su/liberte