Back in December, online backup provider Backblaze presented a simple question: "Enterprise Drives: Fact or Fiction?" The answer, at least based-upon the company's small sampling of drives (over 10,000 "drive hours" were tested), it was discovered that enterprise drives don't necessarily live longer. In fact, in Backblaze's particular case, they lived shorter lives.

Of course, "enterprise" drives are designed for more than just longer life, but this kind of test proves that those who might opt for so-called enterprise / RAID drives might be overspending for features they don't need, with the false assurance of buying drives that will live longer. This discovery struck home with me, because I run 9 hard drives in 24/7 operation, and within one month last summer, two "RAID Edition" 2 TB drives (with 13K hours) died on me. At the same time, I have two consumer drives with 20K+ hours that are still truckin' on.

Nonetheless, while drive reliability is important for a home user, it's imperative to a company like Backblaze which is running thousands of them. A fault at home might result in a dead drive, but here, it could result in dozens or potentially hundreds.

After compiling some data from the 27,000+ consumer drives used in its "storage pads" at the end of 2013, Backblaze has given us its results of which drives have lasted the longest:

Seagate doesn't fare too well here, with an average drive lifespan of 1.4 years. Hitachi (a WD company), which had an almost equal number of drives in use lived quite a bit longer - 2.0 years. The number of WD drives used is paltry in comparison to Seagate and WD, but the result is impressive: 2.5 years. While it's a little unfair to mention Toshiba when a mere 174 drives were used, the result there is no less appalling. Samsung's result is impressive... but 18 drives? That's hardly definitive.

Backblaze goes on to say that there are some hard drives it simply hasn't been able to deploy into its storage pods - and if you're thinking "green", you'd be spot-on. In particular, WD's Green and Seagate's LP might be fine for the home user, but when put into a heavy-load environment like this, errors begin to creep up immediately - which is likely due to vibration.

The below table gives a break-down of the drives Backblaze had in operation between the three leading vendors. Given the number of people who've told me that they've had issues with Seagate's 1.5 TB drives, that result doesn't strike me as too much of a surprise, but when comparing the 3 TB model to WD's 3 TB model, the differences in reliability are stark - so stark, I'd have to imagine that Seagate will feel compelled enough to issue a statement on it.

Looking at the data further, we see that Seagate owns the top 7 spots for the most unreliable drives. Barracuda Green 1.5 TBs had a failure-rate of 120% (they lasted less than a year on average), while the Barracuda 1.5 TBs suffered at a rate of 25.4%. There's a definite trend with the 1.5 TB drives from Seagate, as the third spot also belongs to one such model - Barracuda LP.

WD makes an appearance in slot number 8 with its 1 TB Green drive, which holds a failure-rate of 3.6%. Rounding out the top 10 is WD's 3 TB Red with 3.2%, and Hitachi's 4 TB Deskstar, with 1.5%.

Backblaze goes into some finer detail about its results, but it's clear that Seagate might want to explain itself - the results are simply not in its favor at all. When the data is taken a look at from a 3-year standpoint, Hitachi drives are 96.9% likely to survive the entire time, while WD's drives settle in at 94.8%. Seagate's drives, by contrast, have a 73.5% survival rate.

So, what drives does Backblaze trust enough to continue buying? Interestingly enough, the most common purchase is Seagate's Desktop HDD.15 4 TBs (which, not surprisingly, the company is monitoring closely). In second-place is WD's 3TB Red, although since WD recently issued a 4 TB model, we wouldn't be surprised to see Backblaze begin to consider those.

What are your thoughts on Backblaze's results? Will they influence your next hard drive purchase?