How is fault-tolerant different from redundant?

Answer 1

Although fault tolerance and redundancy both improve overall system reliability, the latter accomplishes that by adding circuit components. Ironically, adding components reduces MTBF (mean time between failures), since the more components a circuit contains the greater the failure rate must be. Think of this way: A power boat with two outboard engines will necessarily have a greater failure rate than a boat with just one engine, because it is twice as likely that an engine will fail. But ask any skipper whether he feels less comfortable about have two engines! Clearly, if a motor fails, you don't have to call Sea-Tow. Fault tolerance is different and a bit harder to define. Circuits can be overspeced and designed to within certain tolerances and accuracies. In a strict sense, when components drift out of those tolerances, the circuit is not performing to spec. You could say that a fault has occurred. But what if the overall system still performs its intended function with no apparent degradation to the end-user? We say those systems are fault tolerant.