A hard disk can suddenly bite the dust thanks to a lightning strike, an accidental drop from a decent height, or an electronic component shorting out. However, most of the time hard disks die a slow death. Along the way, hard disks almost always show some signs of decay, such as the following:

• Spin-up time gradually slows.

• Drive temperature increases.

• The seek error rate increases.

• The read error rate increases.

• The write error rate increases.

• The number of reallocated sectors increases.

• The number of bad sectors increases.

• The cyclic redundancy check (CRC) produces an increasing number of errors.

Other factors that might indicate a potential failure are the number of times that the hard drive has been powered up, the number of hours in use, and the number of times the drive has started and stopped spinning.

Since about 1996, almost all hard-disk manufacturers have built into their drives a system called Self-Monitoring, Analysis, and Reporting Technology, or SMART. This system monitors the parameters just listed (and usually quite a few more highly technical hard disk attributes) and uses a sophisticated algorithm to combine these attributes into a value that represents the overall health of the disk. When that value goes beyond some predetermined threshold, SMART issues an alert that hard-disk failure may be imminent.

Although SMART has been around for a while and is now standard, taking advantage of SMART diagnostics has, until now, required third-party programs. However, Windows Vista comes with a new Diagnostic Policy Service (DPS) that includes a Disk Diagnostics tool that can monitor SMART. If the SMART system reports an error, Vista displays a message that your hard disk is at risk. It also guides you through a backup session to ensure that you don't lose any data before you can have the disk replaced.