Strong knowledge extractors for public-key encryption schemes

Abstract

Completely non-malleable encryption schemes resist attacks which allow an adversary to tamper with both ciphertexts and public keys. In this paper we introduce two extractor-based properties that allow us to gain insight into the design of such schemes and to go beyond known feasibility results in this area. We formalise strong plaintext awareness and secret key awareness and prove their suitability in realising these goals. Strong plaintext awareness imposes that it is infeasible to construct a ciphertext under any public key without knowing the underlying message. Secret key awareness requires it to be infeasible to produce a new public key without knowing a corresponding secret key.