A formal taxonomy of privacy in voting protocols

Abstract

Privacy is one of the main issues in electronic voting. We propose a family of symbolic privacy notions that allows to assess the level of privacy ensured by a voting protocol. Our definitions are applicable to protocols featuring multiple votes per voter and special attack scenarios such as vote-copying or forced abstention. Finally we employ our definitions on several existing voting protocols to show that our model allows to compare different types of protocols based on different techniques, and is suitable for automated verification using existing tools.