Novel model of intermittent faults for reliability and safety measures in long-life computer systems

Abstract There are few evaluation methods which can effectively model intermittent faults, although they account for most of the faults occurring in fault-tolerant computer systems. In this paper we present a general purpose, three-state Markov model which calculates the system reliability and safety for fault-tolerant systems by taking both the effects of permanent and intermittent faults into consideration. We also evaluate the reliability and safety of the duplex system and standby system, both of which are most suitable in long-life unmaintained applications, in the presence of permanent and intermittent faults. From the analysis, we show that the standby system is more reliable than the duplex system and that the safety of the duplex system is superior to that of the standby system. In particular, the simulation results show that a larger permanent-intermittent ratio and/or larger active-inactive ratio make a digital system less reliable and safe.