Last week I presented the work Analyzing the Gold Star Scheme in a Split Tor Network at SecureComm in London:
Tor is an anonymity network and two challenges in Tor are (i) to overcome the scalability problems of Tor’s current network information distribution scheme, and (ii) to motivate users to become operators of nodes. Several solutions have been proposed to address these challenges. We investigate the ramifications of combining two seemingly promising proposals, i.e., splitting the Tor network into several sub-networks (for better scalability), while using the Gold Star scheme (for motivating users to become node operators). Through simulation, we show that the sub-networks are likely to end up in a state of highly imbalanced division of size and bandwidth. This threatens the security and worsens the scalability problem of Tor. We identify the ratio of nodes given a gold star and the fact that a gold star is solely awarded based on a node’s bandwidth, being highly skewed in practice, as two factors that contribute to an imbalanced split. We explore several potential mitigating strategies and discuss their strengths and shortcomings.