Meshkin: A Fault Tolerant Computer Architecture with Distributed Fault Detection and Reconfiguration

1989 
The advent of fly-by-wire aircraft in conjunction with the desire of aircraft suppliers and users to significantly decrease maintenance costs have led to the development of a new set of objectives and requirements for the highly reliable computer systems controlling aircrafts. A new computer system architecture, entitled MeshKin, is presented here. MeshKin has been designed for highly reliable applications such as embedded, real-time avionics electronic systems. The architecture employs quad-modular-redundancy and can sustain at least two independent, non-simultaneous faults; failing in a deterministic fashion upon encountering a third fault. This system diverges from traditional fault tolerant architectures in the placement and management of redundant elements, as there is no one redundancy management module per se. The system’s distributed fault detection, diagnosis, reconfiguration and recovery algorithms are detailed and analyzed, followed by a discussion on issues relevant in the implementation of the architecture.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    10
    References
    4
    Citations
    NaN
    KQI
    []