Autonomous Election Verification with Zeeperio

Abstract

This thesis presents Zeeperio, a verifiable voting protocol that makes end-to-end election verification routine and automatic. Zeeperio uses Eperio’s election table structure and replaces the cut-and-choose auditing with a custom succinct zk-SNARK proof. This proof shows that all ballots are properly accounted for and that the published outcome is correct. The proof is submitted to an Ethereum smart contract for public, automatic verification, so an incorrect tally is rejected without relying on voluntary auditors. We implemented the protocol, and evaluated a 100,000-ballot, 5-candidate election. The proofs resulted in an on-chain verification cost of ~4.2 million gas, which translates to ~$30 USD in fees on Ethereum at the measured prices. These results show that large scale public verification can be inexpensive and compatible with existing paper-election workflows.

Summary for Lay Audience

This thesis is about making election result checking routine, not optional. In many elections, the public must largely trust that the reported totals are correct. Cryptographic end-to-end verifiable voting systems allow for voters and observers to check an election, but the checking process is often dependent on volunteers and specialized tools, so it may not happen consistently. We present Zeeperio, a protocol that produces a short cryptographic proof showing that the published election outcome matches the recorded ballots. Instead of asking people to manually audit large datasets, Zeeperio generates a compact proof that can be verified automatically. The key idea is to have a public smart contract verify the proof on the Ethereum blockchain. If the election data is inconsistent or the tally is wrong, the proof fails and the contract rejects it. This makes verification automatic and publicly visible, while still allowing independent auditors to check the same proof themselves. We implemented Zeeperio and evaluated it on an election with 100,000 ballots and 5 candidates. The proofs are very small and verifying it on Ethereum’s test network cost under $30 USD (at the time). Compared to prior approaches that require very large proofs and heavy computation, these results show that large-scale public verification can be practical and inexpensive.

Description

Keywords

E2E verifiable voting, zk-SNARKs, election verification, Ethereum, smart contracts, cryptographic proofs, Polynomial-IOPs, zero-knowledge proofs

DOI

Collections