If a piece of equipment has an MTBF of 1,000,000 hours' usage, that doesn't mean that any piece of equipment can be expected to last 1,000,000 hours. Rather, it means, roughly, that if 1,000,000 pieces of equipment which are within their rated service lifetime are each operated for one hour, or 100,000 pieces operated for ten hours (but still within rated lifetime), or 60,000,000 for one minute, etc. there will be roughly one failure in the lot. Note that rated service lifetime is an entirely orthogonal to MTBF. Consider the following two types of widgets:
- Every widget, regardless of age, has a 0.1% chance of failing every hour.
- Out of every billion widgets, all but one will operate for precisely 61 minutes and then die; that one will die after 30 minutes; the widgets have a specified service lifetime of 60 minutes.
The first type of widget would have an average lifetime of about 1,000 hours, and also have an MTBF of about 1,000 hours. The second would have an average lifetime of 61 minutes, but an MTBF of 1,000,000,000 hours within its service lifetime. While it may seem odd to say the second device has an MTBF that's almost billion times as long as the expected lifetime, the MTBF is hardly a meaningless figure.
Suppose one is going to conduct an experiment that requires that 1,000,000 devices all work perfectly for an hour, after which they will all be scrapped. If any device fails, the entire experiment will be ruined. Which would be more useful--a device which will last an average of 1,000 hours but has an MTBF of only 1,000 hours, or a device which would last at most 61 minutes, but would have only a one in a billion chance of failing to meet that mark?
3A few things worth noting might be the fact that early failures are often called burn-in. Manufacturers that have much lower early failures often run devices through their burn-in phase. Also that pure electronics do not exhibit a wear out period and only a burn in. – Kortuk – 2011-10-25T10:26:21.690
1Note that when you are calculating the MTTF (or MTBF), you're usually using just a single distribution to model the failures. Therefore the calculation is either based on "infant mortility", "normal life", or "end of life wear-out" distribution. The only thing which distinguishes these three distributions is the Weibull shape parameter, if you're using Weibull as your basic distribution. The only case in which the failures would come out of the "normal life" distribution is when time would have no effect on the failure rate, and therefore the distribution would be exponential. – None – 2011-10-25T13:49:24.883
2MTTF is primarily useful as an indication of what sort of life you should expect from the device or widget. It cannot be, for obvious reasons, an exact prediction of the date of failure of the device. It's only an estimate based on the statistical analysis of the available data and should be considered only as such. Useful for budgeting (how long should I amortize or depreciate the costs here) and planning (how long can we expect the widget to perform before we have to get the next one). – music2myear – 2011-10-25T18:20:41.303
First off, what exactly is a "disk failure"? – Kaitlyn Mcmordie – 2011-10-26T00:16:35.690
2@Kaitlyn - I guess you're referring to bad sectors. I'd say a disk failure is when you can't read from or write to the drive any longer. Usually a mechanical error, like a head crash. This usually happens when you still have plenty of good sectors left. – stevenvh – 2011-10-26T06:55:57.523