Research And Analysis On Leakage-resilient Public Key Encryption Protocols

Traditionally, when proving the security of a cryptographic primitive, it is treated as a black box in which the adversary can learn the input and output, but not get any internal state information. Unfortunately, the security mentioned above does not match the security of real-life implementations if there exists key-leakage attack such as side-channel attack, hacking attack, and more recently “cold-boot” attack, the adversaries can learn partial information on the secret key. The encryption schemes which have been proved security do not retaining security any more against such attack. Hence the mechanism of leakage-resilient cryptography becomes hot topic of modern cryptography. If an encryption scheme is maintain security even when the adversary may additionally learn some partial information of the internal secret state, it is said to be leakage resilient cryptosystems. The research of the public key encryption algorithms against after-the-fact leakage attack was initiated until the notion of entropic security was proposed by Halevi and Lin. It is very important and meaningful to improve the security of public key encryption scheme against after-the-fact leakage attack in the bounded memory leakage model.This paper is intended mainly to research on leakage-resilient public key encryption scheme and its analysis. The specific results follow:1. Inspired by the nation of Halevi and Lin’s entropic security, the nation of security of leakage-resilient public key encryption schemes which are sematic security and chosen-ciphertext attack security against after-the-fact leakage attack is presented. By using extractor and weak hash proof systems, in the bounded memory leakage model we propose two leakage-resilient public key encryption schemes against after-the-fact leakage attack. Two efficient instantiations on the decision version of truncated q-ABDHE assumption are constructed.2. Based on a public key encryption under this bounded memory-leakage model, a new construction of a provably secure three-move leakage-resilient identification scheme is proposed. The proverP can prove its identity to the verifier V with corresponding pk using the secret key sk,. With some optimization of NS encryption scheme, our identification scheme against pre-impersonation leakage attack tolerates leakage up to (12n)L bits which is close to the size of the secret keysk, whereLis the length ofsk. Based on the nation of entropic security, our identification scheme against anytime leakage attack is resilient. The security of this construction can be proved in the standard model under the discrete logarithm assumption.