Abstract
Nanometer-scale CMOS devices, as well as new devices based on quantum technologies are expected to gradually replace current CMOS devices within the next ten to fifteen years. However it is expected that these devices will be prone to failures of several types, until radically new fabrication processes yet to be developed stabilize, and error absorbing methodologies adapted to expected failures are developed and massively applied. This paper proposes a method and the underlying system architecture for improving the fault-tolerance, based on a feed-forward four layer structure which can accommodate deep submicron CMOS circuits and novel quantum structures. Simulation results show a significant improvement of yield with respect to widely applied triple-redundancy and majority voting techniques. A programmable logic array arrangement of the proposed architecture is demonstrated.