Concurrent Fault Detection for Secure QDI Asynchronous Circuits

Asynchronous micropipelined designs are self-checking against permanent stuck-at faults making the circuits highly resistant against invasive fault attacks. A single transient fault, however, can result in a data token insertion or deletion which can cause a continuous stream of erroneous outputs that can be exploited by a malicious attacker. We propose a system level error detection method targeted for detection of faults that cause data token insertions and deletions. The method is based on minimum distance robust error detecting codes and exploits the repeating errors in linear networks to improve the error detection performance.