Reading the original PDF, I noted the following:
"Our testing framework detects unserializable writes
with HDD#1, too. This indicates that some low-end hard
drives may also ignore the flush requests sent from the
kernel. On the other hand, HDD#2 incurred no failures
due to power faults. This suggests that this high-end hard
drive is more reliable in terms of power fault protection."
Now, for ref, the above involved testing TWO spin-type HDD's .. HDD#2 was the 2nd drive of the pair, and it reportedly did not fail.
While I was doing bio research I recall this joke:
"33% of the test subjects improved with the treatment, 33% had no improvement, and the 3rd mouse got away"
WHY oh why even *mention* such a meaningless 'result' in a paper that otherwise seems to adhere to proper scientific methods??
Oh , and yes, of course, how come it's ONLY the researchers who learned which drive brands did well??
Was this gov-funded? This is important, for if we were to learn that the SSD's that did well are no longer available on the market, it would save us a LOT of time, not even considering that any of the 'surviving' SSD's might be known to have other serious flaws in design .. (i.e.: in performance ) !!