Temperature attacks on TRNG and QRNG devices

This research work analyzes the performances of different TRNG and QRNG devices used for the production of random numbers after having subjected them to different low cost attacks. The aim of this thesis is to test the resilience of these devices, i.e. if and how much they are able, even after being attacked, to perform their work correctly. This analysis is performed by means of different statistical tests which measure the behavior of above mentioned devices: beside the most known statistical tests in commerce such as Ent, TestU01, BoolTest and PractRand, the performance comparisons have been analysed also using BitReps, a novel statistical test built by the University of Kent, UK. This being said, the proposed work aims also at analysing the performances of this novel statistical test, which until now had not yet been tested except with test data.