[Paper Review] Dining Cryptographers are Practical

Christian Franck, Jeroen van de Graaf|arXiv (Cornell University)|Feb 10, 2014

Cryptography and Data Security8 references3 citations

TL;DR

This paper proposes a practical extension of the dining cryptographers protocol using Pedersen commitments and zero-knowledge proofs to enable efficient, verifiable collision resolution in unconditionally secure anonymous communication. By binding participants to their secret keys and allowing detection of computationally limited disruptors through cryptographic proofs, the scheme achieves optimal throughput without prior reservation phases, significantly improving practicality while preserving unconditional sender and recipient untraceability.

ABSTRACT

The dining cryptographers protocol provides information-theoretically secure sender and recipient untraceability. However, the protocol is considered to be impractical because a malicious participant may disrupt the communication. We propose an implementation which provides information-theoretical security for senders and recipients, and in which a disruptor with limited computational capabilities can easily be detected.

Motivation & Objective

• To address the impracticality of the original dining cryptographers protocol, where malicious participants who disrupt communication are hard to detect.
• To enable efficient detection of computationally limited disruptors without compromising unconditional anonymity.
• To design a collision resolution mechanism that achieves optimal throughput and avoids lengthy detection phases.
• To integrate cryptographic proofs into the protocol so that message retransmissions can be verified without revealing sender identities.
• To make unconditionally secure anonymous communication viable for real-world applications like electronic voting and low-latency messaging.

Proposed method

• Uses Pedersen commitments to computationally bind participants to their secret keys, ensuring that any deviation from correct behavior can be cryptographically detected.
• Employs zero-knowledge proofs to verify that participants correctly retransmit messages in a tree-based collision resolution protocol without revealing sender identities.
• Constructs verifiable superposed receiving using a binary tree structure where each collision is split into two parts, and cryptographic statements ensure correct retransmission in the right branch.
• Applies homomorphic properties of Pedersen commitments to allow verification of message consistency across multiple retransmissions and to detect inconsistencies indicating cheating.
• Uses a trusted investigator to collect and validate zero-knowledge proofs, only revealing evidence of cheating when necessary to maintain privacy.
• Supports long messages by extending commitments to vectors using multiple generators, enabling secure commitment to multi-element messages.

Experimental results

Research questions

• RQ1 Can a dining cryptographers protocol be made practical while preserving unconditional sender and recipient untraceability?
• RQ2 How can computationally limited disruptors be efficiently detected in an information-theoretically secure protocol?
• RQ3 Is it possible to achieve optimal collision resolution throughput without requiring a reservation phase?
• RQ4 Can zero-knowledge proofs be used to verify message retransmissions without compromising anonymity?
• RQ5 How can key agreement be securely established in a way that supports long-term protocol security and disruptor detection?

Key findings

• The proposed scheme enables verifiable collision resolution with optimal throughput, eliminating the need for prior reservation phases used in earlier protocols.
• Malicious participants who disrupt communication can be detected through zero-knowledge proofs that verify correct message retransmission behavior.
• The use of Pedersen commitments allows computationally binding commitments to secret keys, enabling detection of cheating without revealing sender identities.
• A trusted investigator can efficiently detect and isolate disruptors by validating zero-knowledge proofs, reducing the need for repeated probabilistic retransmissions.
• The protocol supports long messages through vector commitments using multiple generators, maintaining security and efficiency.
• The scheme maintains unconditional anonymity while enabling practical detection of computationally bounded disruptors, marking a significant step toward deployable untraceable communication systems.