Re:Dispelling Backblaze's HDD Reliability Myth
by nwh on Friday January 24, 2014 @08:48PM
(#7119456)
There's a typo in that link.
http://static.googleusercontent.com/media/research.google.co...
I'm more inclined to believe Google's results than anybody elses, the amount of data they have is staggering and the results they got are fairly clear cut. I personally don't bother cooling spinning disks, it just seems like a complete waste of time and effort based on that paper.
Google has a whitepaper[0] on a large deployed sample size of drives which disagree with the author's claims here. I agree that more data from backblaze would be nice, but their claims do seem to have more plausibility than the author suggests. (If he has methodology issues with the Google paper, that would be more interesting)
[0]- http://static.googleusercontent.com/media/research.google.co... (pdf warning)