STATE RESTORATION AT RUNTIME AFTER TRANSIENT HARDWARE-FAULTS IN REDUNDANT REAL-TIME SYSTEMS

Programmable Electronic Systems (PESs) for real-time applications that cannot immediately be transferred to safe states must provide especially high degrees of fault-tolerance. Conventionally, this reqirement is satisfied not only by configuring multiple PESs redundantly but also by applying redundant structures within each PES. Only few systems provide the capability to re-start PES units at runtime. Re-starting units must copy the internal state from their redundant counterparts, and then - when state equivalence has been reached - rejoin redundant processing. This State Restoration at Runtime prevents redundancy attrition due to transient faults, since failed channels can be brought back on line. Here, the associated problems are explained, the advantages and disadvantages of existing techniques are discussed, and a hardware-supported state restoration concept is introduced.