Advanced cryptographic techniques for building verifiable and transparent electronic voting protocols
Electronic voting presents many challenges due to its multiple security requirements. Some of the challenges are related to guaranteeing voters' privacy and system's transparency, which are hard to satisfy simultaneously. Electronic voting also presents other challenges such as usability,...
| Author: | |
|---|---|
| Format: | doctoral thesis |
| Publication Date: | 2017 |
| Country: | España |
| Institution: | Universitat Politècnica de Catalunya (UPC) |
| Repository: | UPCommons. Portal del coneixement obert de la UPC |
| Language: | English |
| OAI Identifier: | oai:upcommons.upc.edu:2117/111503 |
| Online Access: | https://hdl.handle.net/2117/111503 https://dx.doi.org/10.5821/dissertation-2117-111503 |
| Access Level: | Open access |
| Keyword: | Electronic voting Verifiability Cryptography Zero-knowledge proofs Vot electrònic Verificabilitat Criptografia Proves de coneixement zero Àrees temàtiques de la UPC::Matemàtiques i estadística |
| Summary: | Electronic voting presents many challenges due to its multiple security requirements. Some of the challenges are related to guaranteeing voters' privacy and system's transparency, which are hard to satisfy simultaneously. Electronic voting also presents other challenges such as usability, particularly from the voter's side. We study two particular problems of electronic voting. Cast-as-intended verifiability comprises those mechanisms which assure the voter that her cast ballot corresponds to her chosen voting options. Current proposals put the verification burden on the voter, something which is undesirable in real-world elections, where both technically skilled and non-skilled voters participate. In this thesis, we introduce the concept of universal cast-as-intended verifiability, which provides mechanisms which allow any entity to check that any ballot corresponds to the voter's selections - without revealing them. We formally define what universal cast-as-intended verifiability is and we give an electronic voting protocol satisfying this property. The other problem we have studied is the problem of invalid votes in electronic elections. Since a common selling point of electronic voting is that it avoids voters inadvertently spoiling their votes, deliberately spoiled ballots appearing in the tallying phase of an electronic election can cause mistrust on the system. Indeed, election stakeholders might think that the system is flawed or that it was exploited somehow. To avoid this situation, we define the concept of vote validatability, which states the electronic voting system should be able to detect spoiled ballots before they are successfully cast. In addition to formally defining this notion, we design an electronic voting protocol satisfying this property. All these security requirements of electronic voting systems are implemented with cryptographic tools. In addition to encryption and signature schemes, another essential primitive for building electronic voting protocols is zero-knowledge proofs. Zero-knowledge proofs allow a prover to convince a verifier that a statement is true without leaking any other information. These zero-knowledge proofs can be used to, for example, prove that the tally of the election was done properly. Recently, Groth and Sahai constructed efficient non-interactive zero-knowledge proofs for a wide range of statements including, among others, statements appearing in electronic voting. In this thesis we give two contributions on Groth-Sahai proofs. On the one hand, we give a framework for deriving cryptographic assumptions from which to build secure cryptographic protocols. In particular, we build new Groth-Sahai proofs improving the efficiency of currently known constructions. Independently, we show how the original Groth-Sahai proofs can be extended to be compatible with even more statements, how to improve their out-of-the-box efficiency for many of these statements and how to improve their re-usability efficiency among multiple statements. |
|---|