[tor-talk] What was the academic paper in which

[Griffin Boyce]

On 2014-10-19 19:04, Runa A. Sandvik wrote:
> On Sun, Oct 19, 2014 at 11:02 PM, Virgil Griffith <i at virgil.gr> wrote:
>> researchers setup an exit node and then recorded what sites people
>> were going to?
> 
> I believe you are referring to
> http://homes.cs.washington.edu/~yoshi/papers/Tor/PETS2008_37.pdf

   He might also be thinking of the unpublished DOJ study from a couple 
of years ago.  I asked about this back in May, and here was Andrew's 
response:

On 2014-05-25 16:35, Andrew Lewman wrote:
> On Sun, May 25, 2014 at 05:40:51AM -0400, griffin at cryptolab.net wrote
> 0.4K bytes in 0 lines about:
> :   Is there a good reference for the assertion by DOJ that 3% of Tor's
> : traffic is "bad"/used for piracy/etc?  This has been referenced in a 
> few
> : talks, but was just wondering if this is written anywhere that can be 
> easily
> : referenced.
> 
> There was an unpublished study in Nevada by some grad students who 
> setup
> a few malware defense appliances on the end of a tor exit relay. They
> found 3% of the traffic passing through their exit relay was tagged as
> malware, by however the appliance was configured to determine malware
> or not.
> 
> They never published their research because they either couldn't get
> ethics board clearance at their university and/or because of the Univ. 
> of
> Colorado exit relay issue at PETS.
> 
> I read a draft of the paper, which was subsequently pulled from
> publication. I've talked to a few organization who allow Tor exits, but
> track good/bad traffic (by their definition) who say 3% sounds high 
> from
> what they have seen. None of these orgs will go on the record, but they
> are some of the largest social network and ecommerce sites in the 
> world.

Source: 
https://lists.torproject.org/pipermail/tor-talk/2014-May/033061.html

-- 
"I believe that usability is a security concern; systems that do
not pay close attention to the human interaction factors involved
risk failing to provide security by failing to attract users."
~Len Sassaman