Realistic Threats to Self-Enforcing Privacy

Abstract

A recent privacy protocol for secure e-polls aims at ensuring the submitting individuals that the pollster will preserve the privacy of their submitted preferences. Otherwise the individuals can indict the pollster, provided that the pollster participates actively in this phase. The analysis of the protocol in a realistic threat model denounces that a malicious pollster that abuses the private preferences by disclosure will arguably not help out during its own indictment. Therefore, the protocol ensures insufficient fairness among their participants because it gives the pollster some advantage over the individuals. Two variant protocols are introduced and analysed in the same threat model - one is found to move the advantage over the individuals, the other is found to achieve a satisfactory level of fairness.