ParentFull threadpiaste·I assume you take the first batch of N drives, run them for K weeks, see how many fail, and extrapolate.The methodology ought to be published, though.View on HN