Coercion-Resistant Verifiable Web-based Elections in Linear Time

Abstract

Voter coercion broadly includes any attempt to influence the secrecy of a voter’s ballot, including bribery, vote stealing, and forced abstention. This risk, which affects the fundamentals of privacy in a democracy, is especially prevalent in remote voting systems. Numerous coercion-resistant protocols have been proposed that manage to counter this risk, but compromise on time-complexity and practicality.

This thesis proposes a new protocol for a coercion-resistant, verifiable remote voting system that tallies votes in linear time. The protocol is modeled using observational equivalences in Applied Pi Calculus. This model is then verified in ProVerif to determine its compliance with a formal definition of coercion-resistance.

The thesis finally presents a web-based application built using the Django framework. The application implements an election system based on the verified protocol. It provides a user interface to interact with the system and cast votes, while server-side Python scripts create, manage, and tally the election.

The protocol is compared against previous proposals based on the number of modular exponentiations required for a single run of an election. The application is evaluated based on the time taken for the execution of different phases in an election.

The results will lead to implementations of coercion-resistant voting applications that can be deployed in elections where the threat of coercion is considered significant.