Non-malleable Codes from Leakage Resilient Cryptographic Primitives

Non-malleable codes (NMC) are used as a relaxation of error correction and error detection codes to guarantee strong privacy where correctness is not the main concern. Usually, a coding scheme is said to be non-malleable with respect to a class of tampering function if any tampering with the codeword, the underlying function changes the codeword to a completely unrelated one, i.e., \(\bot \) or same, in case of unsuccessful tampering. The real life application of such codeword is to provide security against leakage and tampering attacks on the memory, which is also called active physical attacks or hardware attacks. Standard version of non-malleable codes are used to protect highly sensitive data (i.e., secret key of any cryptographic scheme) on private memory of the device. In literature, leakage resilient authenticated encryptions (AE) are used to design such codeword. We show a generic framework to design leakage resilient authenticated encryption and prove it non-malleable with respect to one-time tampering attack. The instantiation of such codeword is based on leakage resilient IV-based encryption scheme along with leakage resilient CBC-MAC and 1-more weakly extractable leakage-resilient hash function (wECRH). When the tampering experiment of our strong NMC returns \(\bot \) , the security is reduced to the security of authenticated encryption and 1-more weakly extractable leakage-resilient hash function.