2011-11-30: A BaCaTeC request for funding was approved.
The dragon project is a cooperation between UCSB seclab and FAU CS 4 on the topic of anonymous networks.
Anonymous networks provide anonymous message exchange across the Internet. The most widely used system is TOR, which does not scale well because of its centralized design. Some research efforts have been undertaken to create decentralized anonymous networks, with the goal to overcome the scalability problems of centralized protocols. However, so far, no design has been successful.
Decentralized protocol designs typically use a distributed hash table (DHT) to store their application data and to make it available to other nodes in the network. These distributed systems suffer from a variety of problems, weakening the anonymity of the networks. In particular, current designs leak information about the participating nodes' identities, slowly reducing their anonymity. This can be addressed by adjusting the lookup algorithm for stored data. Unfortunately, these new algorithms, while they mitigate the information leakage problem, are no longer reliable with respect to the query results returned. Thus, malicious nodes, participating in the DHT, can provide fake query results to other nodes staging different attacks to weaken their anonymity, e.g. by luring them into a fake network that they fully control and monitor.
The goals of this cooperation project are:
Development of novel and innovative distributed storage design that does not leak information about its participants while providing reliable query results
New insights into how to build and scale anonymous communication systems
New insights into possible pitfalls and fallacies with decentralized anonymity systems
New insights in how to scale decentralized, anonymous networks successfully