There is a pervasive dream about a free Internet which is robust,
fully decentralized yet efficient, and which ensures privacy for all
users. For seven years, the Freenet project has been the most visible
embodiment of this vision. This talk will show that the recent 0.7
release of Freenet -- marketed to solve most of the problems --
entirely fails to deliver.
Freenet 0.7 promises efficient routing in restricted-route networks,
often also called friend-to-friend (F2F) networks or darknets. Our
work shows that a crucial step in the routing protocol can be easily
subverted by an adversary which is no more powerful than any ordinary
node operator. The attack targets a fundamental aspect of the routing
protocol; in particular, it does not rely on minor flaws in the
Freenet implementation and can thus not be easily addressed.
The goal of this talk is not to destroy the dream of a free Internet.
Instead, the talk will educate the audience about pitfalls on the path
to utopia, improving our progress to this shared vision by shining a
light on certain dead ends.