Mean Time between Failures

---

Reliability is expressed in the mean time between failures (MTBF) rating of a disk and can run anywhere from 300,000 hours to over 1 million. This figure does not mean that a disk is expected to last 1 million hours (114 years of 24/7 operation). As explained in the Seagate Technology Paper 338.1, Estimating Drive Reliability in Desktop Computers and Consumer Electronic Systems, the MTBF for a drive is defined as the number of power-on hours (POH) per year divided by the first-year annualized failure rate (AFR). In essence, this is an approximation for small failure rates and is intended to represent only a first-year MTBF. The figure does not indicate how long the drive will last, but it does reflect the odds that it will break down in the first year of operation. The higher the number, the less chance it will fail that year. This is explained in greater detail in Chapter 3.

While in an ideal world you would want to go for the highest possible rating on all factors, in reality tradeoffs are involved. For example, a higher rotational speed means quicker data access, but that speed boost is often achieved by reducing platter size and lowering capacity, so perhaps installing a higher disk cache would be better. Paying to get the highest possible MTBF rating might be a waste of money for a workstation that has all of its critical files stored on a server or for a single disk in a mirrored array that can easily be hot-swapped with a replacement if it goes bad. But, by knowing all the different components that go into disk performance and what effect they have on the overall picture, users can intelligently balance the features to get the best mix for their own operations.

---