MTBF is a measure of the expected (or average) time between failures of a given system. This reflects how frequently the system under consideration fails. Using the MTBF and the MTTR[1] one can estimate the expected availability of the system, and the probability of MultipleSimultaneousFailure[2].
Expected Service Availability = MTBF / (MTBF + MTTR[1])
MTTR[1]
| [1] | http://www.linux-ha.org/MTTR |
| [2] | http://www.linux-ha.org/MultipleSimultaneousFailure |
This information provided courtesy of the Linux-HA project at http://linux-ha.org/