Performance Optimization on Practical Quantum Random Number Generators: Modification on Min-entropy Evaluation and Acceleration on Post Processing

TL;DR

This paper improves practical quantum random number generators by refining min-entropy estimation to ensure security and accelerating post-processing using FFT-based Toeplitz matrix multiplication, optimizing speed and security.

Contribution

It introduces a modified min-entropy evaluation method for laser phase noise-based QRNGs and accelerates post-processing with FFT, enhancing both security and efficiency.

Findings

Modified min-entropy estimation guarantees no information leakage.

FFT-based Toeplitz matrix multiplication reduces post-processing time complexity to O(nlogn).

Optimal block length improves processing speed for fixed-length raw sequences.

Abstract

Quantum random number generation is a technique to generate random numbers by extracting randomness from specific quantum processes. As for practical random number generators, they are required not only to have no information leakage but also have a high speed at generating random sequences. In this paper, we consider the generators based on laser phase noise and propose a method to modify the estimation of min-entropy, which can guarantee no information leakage to the eavesdropper. We also accelerate post processing based on Toeplitz matrix with Fast Fourier Transformation, reducing its time complexity to O(nlogn). Furthermore, we discuss the influence on post processing speed by block length and find a proper block length to process a fixed-length raw sequence.