2019UCL Discovery (University College London)Requires access

Designing efficient zero-knowledge proofs in the ideal linear commitment model

Jonathan Bootle

Open publisher page 0 citations

Abstract

Zero-knowledge proofs are cryptographic protocols where a prover convinces a verifier that a statement is true, without revealing why it is true or leaking any of the prover’s secret information. Since the introduction of zero-knowledge proofs, researchers have found numerous applications to other cryptographic schemes, such as electronic voting, group signatures, and verifiable computation. Zero-knowledge proofs have also become an integral part of blockchain-based cryptocurrencies. Thus, designing efficient zero-knowledge proofs is an important goal. Recently, the design space has become extremely large. To simplify protocol design, designers have begun to separate the process into modular steps. Information theoretic protocols are designed in idealised communication models and compiled into real protocols secure under cryptographic assumptions. In this thesis, we investigate the Ideal Linear Commitment model, which characterises interactive zero-knowledge protocols where the prover and verifier use homomorphic commitment schemes. We demonstrate the model’s power by exhibiting efficient protocols for useful tasks including NP-Complete problems and other more specialised problems. We demonstrate the model’s versatility by compiling the idealised protocols into real protocols under two completely different cryptographic assumptions; the discrete logarithm assumption, and the existence of collision-resistant hash functions. We show that the Ideal Linear Commitment model is a useful and effective abstraction for producing zero-knowledge protocols. Furthermore, by identifying the limitations of the model and finding protocols outside these constraints, we display special techniques which result in more efficient zero-knowledge proofs than ever. The results are novel and highly efficient protocols. Results include the first ever discrete-logarithm argument for general statements with logarithmic communication cost, the first ever three-move discrete-logarithm argument for arithmetic circuit satisfiability with sub-linear communication costs, and an argument for list membership with sub-logarithmic communication, less than the number of bits required to specify a list index. Every single one of our protocols improves the theoretical state-of-the-art.

About this research paper

What this paper is about

Zero-knowledge proofs are cryptographic protocols where a prover convinces a verifier that a statement is true, without revealing why it is true or leaking any of the prover’s secret information. Since the introduction of zero-knowledge proofs, researchers have found numerous applications to other cryptographic schemes, such as electronic voting, group signatures, and verifiable computation. Zero-knowledge proofs have also become an integral part of blockchain-based cryptocurrencies. Thus, designing efficient zero-knowledge proofs is an important goal. Recently, the design space has become extremely large. To simplify protocol design, designers have begun to separate the process into modular steps. Information theoretic protocols are designed in idealised communication models and compiled into real protocols secure under cryptographic assumptions. In this thesis, we investigate the Ideal Linear Commitment model, which characterises interactive zero-knowledge protocols where the prover and verifier use homomorphic commitment schemes. We demonstrate the model’s power by exhibiting efficient protocols for useful tasks including NP-Complete problems and other more specialised problems. We demonstrate the model’s versatility by compiling the idealised protocols into real protocols under two completely different cryptographic assumptions; the discrete logarithm assumption, and the existence of collision-resistant hash functions. We show that the Ideal Linear Commitment model is a useful and effective abstraction for producing zero-knowledge protocols. Furthermore, by identifying the limitations of the model and finding protocols outside these constraints, we display special techniques which result in more efficient zero-knowledge proofs than ever. The results are novel and highly efficient protocols. Results include the first ever discrete-logarithm argument for general statements with logarithmic communication cost, the first ever three-move discrete-logarithm argument for arithmetic circuit satisfiability with sub-linear communication costs, and an argument for list membership with sub-logarithmic communication, less than the number of bits required to specify a list index. Every single one of our protocols improves the theoretical state-of-the-art.

Why it matters

A significance statement is not available in the OpenAlex record.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Zero-knowledge proofs are cryptographic protocols where a prover convinces a verifier that a statement is true, without revealing why it is true or leaking any of the prover’s secret information. Since the introduction of zero-knowledge proofs, researchers have found numerous applications to other cryptographic schemes, such as electronic voting, group signatures, and verifiable computation. Zero-knowledge proofs have also become an integral part of blockchain-based cryptocurrencies. Thus, designing efficient zero-knowledge proofs is an important goal. Recently, the design space has become extremely large. To simplify protocol design, designers have begun to separate the process into modular steps. Information theoretic protocols are designed in idealised communication models and compiled into real protocols secure under cryptographic assumptions. In this thesis, we investigate the Ideal Linear Commitment model, which characterises interactive zero-knowledge protocols where the prover and verifier use homomorphic commitment schemes. We demonstrate the model’s power by exhibiting efficient protocols for useful tasks including NP-Complete problems and other more specialised problems. We demonstrate the model’s versatility by compiling the idealised protocols into real protocols under two completely different cryptographic assumptions; the discrete logarithm assumption, and the existence of collision-resistant hash functions. We show that the Ideal Linear Commitment model is a useful and effective abstraction for producing zero-knowledge protocols. Furthermore, by identifying the limitations of the model and finding protocols outside these constraints, we display special techniques which result in more efficient zero-knowledge proofs than ever. The results are novel and highly efficient protocols. Results include the first ever discrete-logarithm argument for general statements with logarithmic communication cost, the first ever three-move discrete-logarithm argument for arithmetic circuit satisfiability with sub-linear communication costs, and an argument for list membership with sub-logarithmic communication, less than the number of bits required to specify a list index. Every single one of our protocols improves the theoretical state-of-the-art.

Key concepts: Gas meter prover, Zero-knowledge proof, Computer science, Mathematical proof, Cryptographic primitive, Cryptography, Cryptographic protocol, Theoretical computer science

Related papers

Back to paper searchBrowse research topicsOriginal source
Designing efficient zero-knowledge proofs in the ideal linear commitment model — Research Paper | ScholarLens