On the suitability of time-randomized processors for secure and reliable high-performance computing

Abstract

Time-randomized processor (TRP) architectures have been shown as one of the most promising approaches to deal with the overwhelming complexity of the timing analysis of high complex processor architectures for safety-related real-time systems. With TRPs the timing analysis step mainly relies on collecting measurements of the task under analysis rather than on complex timing models of the processor. Additionally, randomization techniques applied in TRPs provide increased reliability and security features. In this thesis, we elaborate on the reliability and security properties of TRPs and the suitability of extending this processor architecture design paradigm to the high-performance computing domain.