Achieving Security Despite Compromise Using Zero-knowledge

Abstract

One of the important challenges when designing and analyzing cryptographic protocols is the enforcement of security properties in the presence of compromised participants. This paper presents a general technique for strengthening cryptographic protocols in order to satisfy authorization policies despite participant compromise. The central idea is to automatically transform the original cryptographic protocols by adding non-interactive zero-knowledge proofs.Each participant proves that the messages sent to the other participants are generated in accordance to the protocol.The zero-knowledge proofs are forwarded to ensure the correct behavior of all participants involved in the protocol, without revealing any secret data.We use an enhanced type system for zero-knowledge to verify that the transformed protocols conform to their authorization policy even if some participants are compromised.Finally, we developed a tool that automatically generates ML implementations of protocols based on zero-knowledge proofs.The protocol transformation, the verification, and the generation of protocol implementations are fully automated.