A fault attack robust TRNG

Abstract

True random number generators (TRNGs) are used for cryptographic operations. By the use of TRNGs secret keys can be generated and the design can be made robust against attacks. In order to simplify further attacks the adversary tries to introduce some bias in the probability distribution of the TRNG output. A possibility to detect such attacks is shown in this paper. In difference to the most former publications all investigations are performed taking into account realistic technology parameters and parasitics.