Hardware for Resilient Computing

ThereHardware for resilient computing is no doubt, system softwareSystem software support of computer resilience is incomplete if hardware does not have essentials required to support it. To demonstrate how our theory of redundancyRedundancy and generalized algorithm of fault toleranceGeneralized algorithm of fault tolerance work in practice we have developed own hardware with required support for primarily we concentrate at fault detection and recoverability of hardware. More detailed description of our hardware design was published in (Blaeser et al. in Evolving systems, 2014; Castano and Schagaev in Resilient computer system design; Schagaev et al in Instruction set and their impact on modern computer architecture, 1989). In our design we address system PRE-requirements (performance, reliabilityReliability and energy) and their trading during functioning of a system when applications need it. In this chapter we simply describe essential elements of the hardware corresponding to the requirements mentioned. Further, we describe hardware solutions that can gain performance and analyze the constraints for a sake of reliabilityReliability. Finally, by creating a reliabilityReliability model for this hardware we demonstrate what kind of reliabilityReliability gain is achievable in our design.