Selene: Voting with Transparent Verification and Coercion Mitigation
- 17:00 3rd March 2016 ( Hilary Term 2016 )Tony Hoare Room, Robert Hooke Building
In conventional cryptographic E2E verification schemes, voters are provided with encrypted ballots that enable them to confirm that their vote is accurately included in the tally. Technically this is very appealing, but voters, election officials etc. need to understand some rather subtle arguments to appreciate the integrity and ballot secrecy guarantees provided by such mechanisms.
A simple way to achieve a degree of verifiability and ballot privacy is to provide each voter with a unique, private tracking number. Votes are posted on a bulletin board in the clear along with their tracking number. Thus voters can visit the WBB confirm that there is an entry with their tracking number showing the correct vote. The beauty of this approach is its simplicity and understandability. There are, however, two drawbacks: we must ensure that trackers are unique and a coercer can demand that the voter reveal her tracking number. It is interesting to note that the coercer must ask for the tracker before posting. If he asks after posting the voter has a simple strategy to fool him: just reads off a tracker number with the coercer's required vote from the WBB.
In this talk, I describe a scheme that addresses both of these problems. The main idea is to close off the coercer's window of opportunity by ensuring that the voters only learn their tracker numbers after votes have been posted. Notification of the trackers must provide high assurance but be deniable. The resulting scheme provides receipt-freeness but also provides a more immediately understandable form of verifiability: voters can find their vote, in the clear, in the tally identified by their secret tracker.
However, there is a sting in the tail: a coerced voter might light on the coercer's tracker, or the coercer may simply claim that the tracker is his. I describe some elaborations of the basic scheme to counter this problem.
Speaker bio
Peter Ryan is Professor of Applied Security at the University of Luxembourg since Feb 2009. He has around 25 years of experience in cryptography, information assurance and formal verification. He was a pioneer of the application of process calculi to modelling and analysis of secure systems, in particular to the characterization of information flows (non-interference) and the application of process algebra (CSP) and model-checking tools (FDR) to the analysis of security protocols. He has published extensively on cryptographic protocols, security policies, mathematical models of computer security and, most recently, end-to-end verifiable election systems. He is the creator of the (polling station) Prêt à Voter and, with V. Teague, the (internet) Pretty Good Democracy verifiable voting schemes. He was also co-designer of the vVote system, based on Prêt à Voter that was used successfully for Lower House elections in Victoria State, Australia in November 2014. With Feng Hao, he also developed the OpenVote boardroom voting scheme and the J-PAKE password based authenticated key establishment protocol. Recently he co-authored the “Remark!” scheme for secure, anonymous marking of exams and has been working on Quantum Authenticated Key Establishment Protocols.Share this:
