We often assume that SSDs are more reliable and long-lasting than HDDs, which makes sense, given that SSDs don’t contain any moving parts. But a new Backblaze report calls that assumption into question. According to the cloud storage company’s real-world tests, SSDs and HDDs are equally reliable.
The data presented in this report is valuable, but it isn’t conclusive. It doesn’t really prove anything. Still, many publications and forums are pointing to it as scripture, so it’s time to read between the lines a bit.
Random nerds on the internet can’t test the reliability, durability, or longevity of HDDs and SSDs. That’s because these sorts of tests require a ton of data. You need to test thousands of drives over a long period of time to find their average failure rate or lifespan, as no two drives are truly identical.
That said, Backblaze is a cloud storage company. It runs through thousands of drives each year, and it publishes quarterly and annual reports on HDD failure rates. Therefore, Backblaze’s data is much more useful than anecdotes from weird Reddit nerds—it provides a clear picture of which drives may work best in a home or business environment.
The most valuable measurement in Backblaze’s report, at least for our purposes, is the AFR or Annualized Failure Rate of SSDs and HDDs. The Annualized Failure Rate shows what percent of SSDs failed throughout 2021. (Notably, this measurement accounts for Backblaze’s testing environment, in which drives are added to and removed from servers throughout the year.)
So, what does Backblaze data say about SSD reliability? According to the company’s report, the AFR or Annualized Failure Rate of SSDs is comparable to that of HDDs.
Nearly all of the SSDs tested by Backblaze have an AFR of less than 1%, with some models falling below the 0.6% AFR that Backblaze looks for in its “most reliable” drives. Not only that, but the cumulative AFR of SSDs (from 2018 to 2021) is around 1.07%, which is actually less than the 1.40% cumulative AFR of HDDs (from 2013 to 2021).
It’s easy to walk away from this data thinking that HDDs and SSDs are equally reliable. And hey, that may be the case! But this data is not conclusive, and Backblaze notes as much in its report.
Backblaze publishes quarterly and annual reports on HDD performance, but oddly enough, this is the first time the company’s shared data on SSDs. And given some of the notes in Backblaze’s report, I’m not surprised that it waited so long to publish this kind of information.
First of all, Backblaze only uses SSDs as boot drives in its storage servers. These drives do more than just boot the servers, and of course, they read, write, and delete files every day. But technically speaking, they are not performing the same tasks as Backblaze’s HDDs—this is a variable that could impact the SSDs’ failure rate. (That said, the impact should be marginal at best.)
Backblaze only began using SSD boot drives in 2018, and that presents another problem. Long-term data is incredibly useful when gauging drive reliability, but most of the SSDs in Backblaze’s servers are relatively new. Notably, long-term data has actually improved the cumulative AFR of HDDs in Backblaze tests.
But here’s the most important point; Backblaze uses far fewer SSDs than HDDs in its servers. The company only tested 2,200 SSDs throughout 2021, but it tested at least 203,168 HDDs in the same year. And even if these numbers were identical, Backblaze has tested HDDs since 2013, so comparing the cumulative AFR of SSDs and HDDs doesn’t make much sense.
We often look at Backblaze’s reports to see which HDDs are the most reliable. It’s a pretty simple idea—if one model of HDD in Backblaze’s servers has a very low AFR, then it’s probably an ideal candidate for your PC or Plex server.
But Backblaze hasn’t tested many models of SSDs. There aren’t any Samsung, PNY, or Western Digital drives in this report, for example. As for the drives that Backblaze actually tested … well, the data may be a bit skewed.
Let’s say you take a quick glance at Backblaze’s report to find which SSDs you should purchase. You may notice that the Seagate ZA2000CM10002 has an AFR of 28%, which is a sign that you should avoid it like the plague, right? Here’s the problem; Backblaze has only tested four Seagate ZA2000CM10002 drives.
Looking to this report to try and find the most reliable SSD is a terrible idea. There just isn’t enough data to draw a serious conclusion.
That said, this Backblaze report is still quite valuable. It’s flawed, for sure, but it’s just the first in several SSD reports that Backblaze will publish. As the company accumulates more data, its reports will paint a picture of how SSDs compare to HDDs, and of course, which SSDs are the most reliable.
I should note that SSDs and HDDs are very different, and that the data presented by Backblaze shouldn’t sway your shopping decisions too much. The average external SSD is often a better option than an external HDD, for example, as it’s smaller and offers a lot more impact resistance.