Failure Trends in a Large Disk Drive Population

The paper presents failure patterns of hard drives collected from a large sample (over 100k hard drives) in Google’s data center. The paper makes the argument that failures are hard to define because failure can be context dependent (e.g. bad driver). Thus, they roughly define a hard drive to have failed if replaced in a repair procedure. As expected, failure results are significantly affected by model/manufacturer, the paper mentions this but does not show such a breakdown.

I found interesting the survival probability breakdown by age after a scan error and after a reallocation and the failure rate by age. I also liked the tracing/detection mechanism (using the Google stack) which is quite general.

I think these results may be interesting for quite some time since such detailed studies for large populations are hard to achieve and not often published, however, at least to me, a breakdown for model/manufacturer before drawing all of these charts would have made much more sense.