Expectations set incorrectly?

So I wonder if we should be thinking that SSDs shouldn’t fail as often as spinning disk. That is, SSDs don’t have moving parts, and so are much less subject to mechanical wear and tear as they are used. But they do fail.

Every brand of SSDs we have used, every one, including Intel, Corsair, RiData, Mushkin … every one, we have seen failures. Some have been absolutely ridiculous in scope (Corsair), some have been mostly due to changes in their mechanical design (RiData) as well as unit failures. Some have just died (Intel, RiData, and Mushkin) without really providing a reason.

So … should we expect something with fewer moving parts to fail less frequently? Or is this simply a guess, and a bad one at that?

There is the possibility that some SATA controllers are incompatible with SSDs over long stretches, in that the normal drives they interact with might not be as sensitive to their signalling as SSDs are. There might be significant SMART issues, such that an incorrect implementation somewhere leads to a shortened life. There may be crappy SSD controller chips that fail enmasse (Corsair).

My expectations had been that they should be more reliable than spinning rust. Maybe the 20+ years of spinning rust … the folks building them, have done a somewhat better job of building quality parts than the plethora of SSD builders offering units?

I dunno. Just doing some introspection.

Viewed 14000 times by 3273 viewers

One thought on “Expectations set incorrectly?

  1. I remember back in the dark past of my engineering degree we did a mechanical reliability course, cycles to failure, revolutions to failure, stress cycles, wear etc etc. I remember the prof saying (in 1988) that we weren’t as good at predicting small-scale electronics failures because we didn’t know everything that caused them to fail, other than obvious things like heat or shorting.

    We always make the assumption that no moving parts = high MTBF. SSDs have had such a rapid development and deployment curve, if they’re all tanking at some point then obviously there’s something the manufacturers don’t have a handle on yet. When you consider the MTBF for spinning rust storage, they’re amazing.

Comments are closed.