Dual versus triplication reliability estimations

The system mean time to failure (MTTF), a measure of reliability, depends largely on the failure rate, the mean time to repair, and the coverage factor. All contribute to the calculation of MTTF and influence the system design in terms of reliability objectives. In dynamically redundant systems, such as dual systems, coverage plays an extremely important role. It is a probability that the system will recover from a single fault. This involves both hardware and software components to properly detect, isolate, and reconfigure around a faulty module without adversely affecting the system operation. A small reduction in coverage can affect the MTTF dramatically. In a static redundant triple modular redundancy (TMR) system, the inherent error detection and correction property gives a near-perfect coverage for the first detected fault. As a result, the calculated MTTF of a TMR configuration is substantially greater than that of a dual system. This paper discusses and compares the reliability estimation of dual and TMR systems.