Continuous Version of Non-malleable Codes from Authenticated Encryption

Non-malleable codes are designed to provide security of highly sensitive data against tampering attacks where traditional error correction and error detection codes fail. An attacker can perform tampering experiment on the codeword but non-malleability property ensures that outcome is either completely unrelated to the original data or the original message, in case of unsuccessful tampering, i.e., tampering has no effect on the codeword at all. Usually, standard non-malleable codes provide security against one-time tampering attack. In literature, it is shown that authenticated encryption can be used in the design of such codeword [22, 31]. The security of such construction breaks when an adversary tampers the codeword more than once. To overcome the situation, continuously non-malleable codes are proposed where an adversary is able to tamper the codeword for polynomial number of times and non-malleability property is preserved. We show a computationally secure construction of continuously non-malleable code from encrypt then MAC based authenticated encryption in 2-split-state model. Earlier codewords are designed using heavy cryptographic primitives like non-interactive zero knowledge proof (NIZK). Our construction is based on non-malleable non-interactive commitment scheme of [32] along with authenticated encryption only. This is the first construction that achieves strong continuous non-malleability without using NIZK, but only relying on non-malleable non-interactive commitment, authenticated encryption and leakage resilient storage. Whenever the tampering experiment triggers self-destruct, the security of continuously non-malleable code is reduced to the security of underlying leakage resilient storage.