COSIC Seminar - Entropy Sources For Cryptographic Random Number Generation (John Kelsey)

Random number generation underlies all of cryptography—if you can’t generate good random numbers, you probably can’t do any useful crypto. In this tutorial, I will go over how cryptographic random number generation works, and then zoom in on entropy sources—the ultimate source of unpredictability in any cryptographic RNG. I’ll discuss the problems of designing and analyzing an entropy source, and the approach we’ve used in SP 800-90B for specifying how they should work and how labs should try to validate them. I’ll also talk about the related problem of extractors, the functions that process entropy-bearing inputs and yield some kind of seed for a deterministic RNG.