Robust Random Number Generation Based on An All-Silicon Approach
Abstract
Secure communication is of great significance in modern world of today. Privacy of personal information is under constant attack of hackers who try to access people's secret messages, social media accounts and bank accounts. Random numbers are used excessively in different areas especially crytography to produce gryptographic keys. Mathematical algorithms generate long sequences of random numbers. However, their security is dubious since when the seed is known, all the numbers in the sequence can be constructed. The most promising way to produce high quality random numbers is to exploit quantum physics with the inherent randomness and unpredictability in quantum phenomenon. We propose a robust methodology for random numbrer generation based on silicon nanocrystals (Si-NCs) LED coupled to a silicon single photon avalanche photodiode (SPAD). Long sequences of data are acquired under different conditions and analyzed by statistical tests to check the robustness and the quality of randomness. All the statistical tests in NIST tests suite are passed successfully without the application of a post processing operation.