“…They argue that for safety-critical systems, reducing this probability to an unknown level is inadequate [8]. Ademaj, Sivencrona, Bauer, and Torin give results from fault injection experiments designed to estimate the percentage of certain Byzantine faults (Slightly Off Specification faults and general asymmetric faults) in the Time Triggered Architecture [1]. They produced Byzantine faults in a bus topology (providing some data about the Byzantine fault rate), but did not observe any in a star topology which employs additional fault protection.…”