MaximumPC 2006 10

r & d BREAKING DOWN TECH —PRESENT AND FUTURE

72 MAXIMUMPC OCTOBER 2006

F

ew situations in life suck as much as

losing a hard drive, especially if you

haven’t backed up the data stored on

it. The experience can be so painful and

upsetting that the enraged victim swears off

that brand of hard drive forever, denounc-

ing that company’s products as unreliable

and untrustworthy.

Hard-drive manufacturers are acutely

aware of this phenomenon and the threat it

poses to their business, and they become

especially conscious when the customer in

question is a massive corporation with mis-

sion-critical data at stake.

But everyone knows that hard drive

failure is inevitable: These are mechanical

devices operating in a sealed environment;

you can’t just pop up the hood on one and

give it an periodic tune-up. Whether the

cause of death is old age (the spindle-motor

lubricant will dry up eventually, and bearings

don’t last forever) or—shudder—a manufac-

turing defect, no drive can live forever.

Because hard-drive manufacturers

can’t prevent eventual drive failure, they’ve

developed self-monitoring, analysis, and

reporting technology (SMART), an early-

warning system that can at least help you

avoid catastrophic data loss. SMART,

which has been integrated into hard drives

since 1996, continually monitors the

drives, on constant look-out for signs of

trouble. It then reports to you, letting you

know either that everything is OK or that

the time has come to really back up your

data and go drive shopping.

PREDICTING FAILURE

Some hard-drive failures are predictable;

others aren’t. Unpredictable failures—sud-

den physical trauma, for instance, or a

lethal power surge—are largely beyond

manufacturers’ control. Predictable fail-

ures, on the other hand, are much easier

to manage. These “slow-death” malfunc-

tions are typically caused by the gradual

degradation of a drive’s components, and

they account for the vast majority of hard-

drive breakdowns.

During the development of SMART,

drive manufacturers performed autopsies

on scores of failed drives returned from

customers. Just as in a human autopsy, they

were seeking to determine the exact cause

of death. Once they’d accumulated a statis-

tically meaningful number of these examina-

tions, commonalities began to emerge; they

detected tell-tale symptoms that proved to

be precursors to crashes. And they estab-

lished thresholds that, once crossed, all but

guaranteed a drive was headed toward that

great trash can in the sky.

In many ways, SMART is like the oil

trouble light on your car’s dashboard. If

your engine doesn’t have enough oil to

keep it lubricated, it will eventually seize

up and destroy itself. This trouble light is

connected to a device that monitors the oil

pressure inside the engine. If the oil pres-

sure drops below a predefi ned level, the

trouble light will glow to alert you.

The only difference between that

trouble light and the SMART sensor in your

hard drive is the fact that you don’t have to

buy a new engine if the oil light comes on—

you just need to add more oil. There’s no

such thing as a friendly SMART warning; if

any of those thresholds are exceeded, your

drive is dying and should be replaced. The

fact that SMART attributes vary not only

from one drive manufacturer to the next,

but even from one drive model to the next,

doesn’t undercut SMART’s value; nor does

the fact that none of the drive manufactur-

ers have agreed on a single defi nitive list of

SMART attributes.

HOW IT WORKS

SMART continuously examines an operating

hard drive, evaluating parameters such as

how long it takes to spin up, the temperature

of the drive, the precise distance between its

read/write heads and its spinning platters,

and more. If SMART determines that one of

these conditions has exceeded the manu-

facturer’s established threshold, it will imme-

White Paper: SMART Hard Drives

Like a canary in a coal

mine, the Self-Monitoring,

Analysis, and Reporting

Technology (SMART) in today’s

hard drives can warn you

of impending doom. Here’s

how it works.

SMART—self-monitoring, analysis, and reporting technology—enables your hard drive to assess its own health and relay that informa-

tion to you. The drive constantly evaluates its key attributes, comparing its state of being to a set of optimum values. If any aspect of its

performance falls outside the manufacturer’s established tolerance range, the drive sends a message to the host controller. The host

controller, in turn, notifi es you during POST on your next restart (or within Windows, if you’re running a SMART utility, such as Speedfan).

HOW IT WORKS SMART in action

The hard drive constantly

monitors itself

BY JOSH NOREM

If your hard drive is failing, SMART

might issue a warning such as this one

during POST. Press F1 and back up as

soon as possible.

If it spots trouble, it sends a message

to the chipset

Windows utilities will notify you with

a “SMART Status Bad” message.

MONITORING ANALYSIS REPORTING