Injection technologies for fault attacks on microprocessors

Abstract

The dependability of computing systems running cryptographic primitives is a critical factor for evaluating the practical security of any cryptographic scheme. Indeed, the observation of erroneous results produced by a computing device after the artificial injection of transient faults is one of the most effective side-channel attacks. This chapter reviews the (semi-)invasive fault injection techniques that have been successfully used to recover the secret parameters of a cryptographic component. Subsequently, a complete characterization of the fault model derived from the constant underfeeding of a general-purpose microprocessor is described, in order to infer how the faulty behavior causes exploitable software errors.