SSD/HDD failure rates

[Voo]

As seen in Anands Intel 510 review, quite interesting numbers:

- Maxtor 1,04% (contre 1,73%)
- Western Digital 1,45% (contre 0,99%)
- Seagate 2,13% (contre 2,58%)
- Samsung 2,47% (contre 1,93%)
- Hitachi 3,39% (contre 0,92%)

- Intel 0,59% - Intel 0.59%
- Corsair 2,17% - Corsair 2.17%
- Crucial 2,25% - Crucial 2.25%
- Kingston 2,39% - Kingston 2.39%
- OCZ 2,93% - OCZ 2.93%

Quite interesting, I think. Although the almost 10% for the 2tb blacks are really something. Although that's probably more influenced by 7.2krpm, which just shows that lots of fast spinning, high density platters aren't a good idea if you want reliable drives.

original
translated

 

[Diogenes2]

Those numbers are hardly scientific.. They are return rates for an e-tailer ... Nothing that shows that they were verified hardware failures, as opposed to user failure or other factors.

It might mean nothing more than Intel buyers returned fewer drives, for whatever reason ..

 

[Voo]

Those numbers are hardly scientific.. They are return rates for an e-tailer ... Nothing that shows that they were verified hardware failures, as opposed to user failure or other factors.

It might mean nothing more than Intel buyers returned fewer drives, for whatever reason ..

So you're claiming that Intel (e.g.) drives get returned for user failures or other reasons more often, because..? If you take two pretty independent facts and claim they influence each other somehow you should have at least a good theory for it, because I don't see why the average user should treat one hard drive different than any other just because there's a different sticker on it.

 

[Diogenes2]

Actually I'm saying the Intel returns, were more likely hardware failures, than were the others.

It's not that the average user treats the drive different, it's a matter of what kind of user buys a particular drive, and why they might be more or less likely to return it.

One possibility could be, that people who pay more for an Intel drive, may be more likely to have more success in installing it correctly and using it properly ..

It would be interesting to see sales volumes versus return rates ..

Chances are, the drives with the highest failure rates, are the ones they sold the most of, and offered more opportunities for someone to screw up, or decide they don't like the way it sounds..

The main problem with the report, is that it tells us nothing about how the returned drives were evaluated, and that they were in fact hardware failures..

 

[Voo]

Chances are, the drives with the highest failure rates, are the ones they sold the most of, and offered more opportunities for someone to screw up, or decide they don't like the way it sounds..

That doesn't make any sense, from my perspective. If the volume is large enough to be statistically significant (which it seems to be), then just selling more only narrows the statistical error down but doesn't mean the return rate goes up for whatever reason (or if it does, why? Why would I be more likely to kill a drive if 100 persons bought it instead of only 10?)

Also Intel SSDs aren't more expensive than other drives out there. Further if we look at Kingston (only sells cheap drives) vs. OCZ (several more expensive ones) the Kingston return rates are lower.


And if we ignore SSDs for a second and look at HDDs we see that there are large differences even though the differences between the different manufacterers HDD's are minimal at best.
Although one thing there is, that we'd have to look what releases fell into the given timespan - if WD was the only one to release a potentially high failure drive (e.g. 7.2k rpm 2tb) and others didn't that would skew the result - not a problem for SSDs.