An Architecture for Runtime State Restoration after Transient Hardware-Faults in Redundant Real-Time Systems

Employing programmable electronic systems (PESs) in safety-critical real-time applications that cannot immediately be transferred to safe states requires especially high degrees of fault-tolerance. Conventionally, this demand is satisfied not only by configuring multiple PESs redundantly, but also by applying redundant processing structures inside each PES. Instead, it is also desirable to provide the capability to rehabilitate a PES´s faulty state by copying the internal state from its redundant counterparts at runtime. Thus, redundancy attrition due to transient faults is prevented, since failed channels can be brought back on line. Here, the problems concerned with state restoration at runtime are stated, the advantages and disadvantages of existing techniques are discussed, and a hardware-supported concept is introduced