Theory and Practice of Secure E-Voting Systems

Abstract

Electronic voting is a popular application of cryptographic and network techniques to e-government. Most of the existing e-voting schemes can be classified into two categories: homomorphic voting and shuffling-based voting. In a homomorphic voting, an encryption algorithm with special homomorphic property (e.g. ElGamal encryption or Paillier encryption) is employed to encrypt the votes such that the sum of the votes can be recovered without decrypting any single vote. An advantage of homomorphic voting is efficient tallying. Tallying in homomorphic voting only costs one single decryption operation for each candidate. In this chapter, the existing e-voting solutions in both categories are surveyed and analysed. The key security properties in both categories are presented and then the existing e-voting schemes in each category are checked against the corresponding security properties. Security and efficiency of the schemes are analysed and the strongest security and highest efficiency achievable in each category is estimated. Problems and concerns about the existing solutions including vulnerability to malicious voters and (or) talliers, possible failure of complete correctness, imperfect privacy, dependence on computational assumptions, and exaggerated efficiency are addressed. New approaches will be proposed in both kinds of solutions to overcome the existing drawbacks in them. In homomorphic e-voting, the authors deal with possibly malicious voters and aim at efficient vote validity check to achieve strong and formally provable soundness and privacy. It can be implemented through new zero knowledge proof techniques, which are both efficient and provably secure. In mix-network-based e-voting, the authors deal with possibly deviating operations of both voters and talliers and aim at efficient proof of validity of shuffling, which guarantees the desired security properties and prevent attacks from malicious participants. It can be based on inspiring linear algebra knowledge and the new zero knowledge proof of existence of secret permutation.