Learn, hack!

URL

http://events.ccc.de/congress/2005/fahrplan/attachments/544-Slides_CovertCommunicationInADarkNetwork.pdf

File name

544-Slides_CovertCommunicationInADarkNetwork.pdf

File size

908.9 KB

MD5

4c96b654af682b562f536660181f114e

SHA1

7318d7c76e4a1876599a109d4b38f3e2c82b26f1

At Defcon 13, we presented our ideas for creating scalable networks where only trusted friends speak directly to each other. In this talk, we will elaborate on this, discussing further experiments and results, as well as our attempts to build such a network for real. We believe that such networks, because they are covert and difficult to detect, are the only viable future for peer-to-peer communication in a time when it is under attack from all sides. The problem with such "dark" networks, or Darknets, is that peers in general are connected only through sequences of friends knowing friends. In order to create a network that still allows global communication, one must find a way of searching efficiently for such paths between hosts. This is where we have focused our attention: if one can efficiently find such paths, then one can use the known techniques of distributed hashtables to make a working, searchable, network. We have now begun our plan towards actually deploying a version of Freenet based on these principles: a file sharing network where only once trusted friends will know that one is even participating. Since it is to be the next version of Freenet, we also intend for it to offer anonymity and replication to protect against censorship. In this talk, we will further elaborate on the ideas and the issues we are facing. We will discuss further experiments and results, as well as our initial experiences of attempting to create such a network for the real world. As before, we will present both the theoretical aspects of our research, which has its roots in deep mathematical results, and the practical aspects of what we are trying to achieve.