Randomness extraction techniques for jittery oscillators

Random numbers represent one of the most sensitive parts of cryptography, being used in different encryption algorithms and/or communication protocols for their property of being irreversible and non-deterministic. Usually, for an attacker, it is almost impossible to break by brute force the encryption algorithm of a secure communication. This is why they focus on breaking some of the consisting components, such as the Random Number Generator (RNG). The basic property of a RNG is that its output depends just on a physical phenomenon and not on anything else (like a previous state). Using a poor RNG (a deterministic one) can lead to communication compromising. A RNG is basically based on a physical phenomenon that cannot be controlled like the jitter from oscillators. This paper describes analyses and demonstrates the techniques that can be used for extracting randomness from jittery oscillators.