Cryptographic Pseudorandom Noise Generators for Lattice-based Cryptography and Differential Privacy

The classical use of a pseudorandom number gen-erator (PRNG) is in the stream cipher encryption where the PRNG is the core for the security of the encryption scheme. The use of PRNGs is also in modern security and privacy techniques such as fully homomorphic encryption (FHE) and differential privacy (DP). When such techniques are employed in many real-world applications such as traditional analytics computing, embedded systems, smart metering and mobile sys-tems, generating a large amount of random numbers and random noises following certain distributions is challenging. In this paper, we propose a construction of a cryptographic pseudorandom noise generator (CPNG) that can also be used for resource-constrained applications. As two use cases, we show how the CPNG can generate Laplace and Gaussian noises for FHE and DP applications. Experimental results on the performance of the proposed generators are presented, along with a comparison.