Study: Hard Drive MTBF Ratings Highly Exaggerated

A study by Carnegie Mellon University, where the researchers took roughly 100,000 hard drives from a variety of manufacturers, concluded that hard drive manufacturers were exaggerating their mean-time before failure (MTBF) ratings on hard drives – failure rates were, on average, 15 times the rated MTBFs. The drives were tested in various operating conditions, including real world scenarios. Researchers found that drive operating temperatures had little to no effect on failure rates: a cool hard drive survived no longer than one running hot. The types of drives used in the study ranged from Serial ATA drives, SCSI and even high-end fiber-channel (FC) drives. Carnegie researchers found that high-end drives did not outlast their mainstream counterparts: "In our data sets, the replacement rates of SATA disks are not worse than the replacement rates of SCSI or FC disks. This may indicate that disk-independent factors, such as operating conditions, usage and environmental factors affect replacement rates more than component specific factors."

The study also found that the number one cause of drive failures was age. Drives tended to start showing signs of failure after roughly five to seven years of service, after which there was a significant increase in average failure rates (AFR). The failure rates of drives that were in their first year of service or shorter was just as high as those after the seven year mark. The average replacement rate of drives ranged from 2-13% annually, indicating that there is a need for manufacturers to re-evaluate the way a MTBF rating is generated. Worst of all, these rates were for drives with MTBF ratings between 1 million and 1.5 million hours. Carnegie researchers concluded that backup measures are a necessity with critically important data, no matter what kind of hard drive is being used.

News source: DailyTech

Report a problem with article
Previous Story

HD Inque Preview

Next Story

Speeding YouTube motorcyclist escapes prosecution


Commenting is disabled on this article.

Wasn’t this the same study posted by Google engineers a short time ago?

In my experience it seems that newer 80Gig IDE and larger hard drives seems to fail far quicker that the old 20 or 40Gig HDD’s – I don’t also think it makes much difference it it’s IDE or SATA as mechanically they are the same. I also have so far yet to see one of our SCSI HDD’s fail. They cost an arm and a leg each (you could buy 4 SATA hard drives with the price difference) but the build quality is far superior to any IDE/ SATA. For a mission critical system this is most important.

I saw SATA, SCSI, and FC. Did they look at any IDE? I know IDE are not the main drive anymore, but there are a lot of them still out there.

The problem with such tests is that they can't measure as well the failure rate of newer hard drives.
A couple of years ago, I had 80GB and 40GB Western Digital drives, and I had to replace them every couple of months. Now I have a WD5000KS (500GB) and WD3200KS (320GB) for quite some time, and they never failed, and the performance is a lot better.
I wonder if WD improved over the years, or is it the better quality of the KS series.

God damn, a study like this and they don't give the brands and their rates... Or at least IF brands had a statistically measurable relationship to failures, in case they wish to hide them for whatever reason. Grrr!

I wish there was a big study done for that... You often can't just rely on forums, because the number of posts for a brands failure will be related to how common the brand is among the users, not just how large ratio of failures it has. Only a study with averages calculated would help there.

on my home server i got a hard drive that was manufactured in 1998. (10 GB IDE Hard Drive) 7200rpm.

it's been running constantly since the year 2000 when i moved the file server computer into a closet. it's been running ever since 24/7/365. only down time was last year when i added some more hard drives and took out another hard drive from that computer. but the whole downtime was less than 20 minutes. the SMART monitor says that the fitness and performance bars are at 80% or higher for this hard drive that i am talking about.

So being that the hard drive is 9 years old. but been running for the past 7 years. it may be to the fact that it isn;t used much, i only have windows on that hard drive. the other hard drives on this computer are used the most. (750 GB total size spanned across 3 hard drives). but these are newer drives. that i added last year only. SMART says that they are 99% fit and at 99% performance efficiency.

Carnegie researchers concluded that backup measures are a necessity with critically important data, no matter what kind of hard drive is being used.

ROFL! Seriously, how much did this study cost for them to make that kind of conclusion? For $10 I could have said the EXACT same thing. Dee dee dee!

Seems reasonable. So try to replace your disks after ~5 years or so. Most likely that will be a good time to change disks anyway, since you probably will get more storage space per $.

so in other words to sum up what that said... heat aint a factor... and drives that are 5-7years old are quite a bit more likely to fail in general... and hard drives that are within 1 year from date of use are likely to fail just like the 5-7 year old ones... so in other words if your drive lasts more than a year odds are it will last a while from what i could understand from what it was saying ... in that case then i got to march of 2007 and dec 2007 before my current hdd's are 1 year old.

bangbang023 said,
And? These are different stories entirely.

They are? Subtitute "Google" for "Carnegie Mellon University" and they're almost identical...