Suppressing Kiddie porn trade / beastiality using "Tar Pit" method
 

Hi.

I was annoyed to find a lot of illegal material on the Gnutella net,
even when just searching for "ogg", trying to find Ogg Vorbis files.

In this context, I don't define "illegal" in a sense of "copyrighted". Here I mean it is simply wrong, immoral and criminal. I am talking about child pornography and beastiality movies. Often these files are tagged by words like "reelkidporn, kporn, beastiality" to stress the originality of the material and make them more "attractive" to potential downloaders.

Not long ago I read about a tool called LaBrea (http://www.hackbusters.net/LaBrea) which was designed to capture incoming TCP connections caused by worm/virus scanning activity. The idea is to fool the scanning host into a nearly endless negotiation/retransmit sequence. The bandwidth consumtion is minimal, while the scanning hosts virtually gets trapped and its potential for causing further harm is minimized.

My idea is to actively suppress filesharing of illegal material by applying tarpit-like methods in the Gnutella network. Even if filesharing of illegal material will never be stopped completely, it will be slowed down and complicated.

The concept is to block outgoing upload ports of Gnutella clients offering illegal material by applying tarpit-like mechanisms. That is:

- establish a TCP connection
- perform the required handshake for the Gnutella download protocol
- and throttle the data transfer to an absolute minimum by forcing resends on the TCP protocol level - but always avoiding a timeout in order to keep the connection up.

The idea is to scan the Gnutella network for illegal content (filename/filesize based) and block any hosts offering such content by jamming their upload slots (any Gnutella host has a limited number of upload slots, as far as I know).

What do you think of such an idea?

Would you consider this a dangerous idea because it could have a negative impact on the technical integrity of the Gnutella filesharing network?

Or do you think the benefits of such an approach would outweigh the potential problems it could cause?

Who would be willing to start an open source project with this "mission" or objective?

Who would actually install this tarpit on his machine, given that the bandwidth consumption is within acceptable limits?

Personal feedback (no hate mail please) to cbuchner1@surfeu.de since I usually do not check this forum very often.