Pure Storage: A must read article

What’s the failure rate of consumer grade flash memory?

Or, to get to the point, what is the real difference between consumer grade and enterprise grade?

This https://storageswiss.com/2015/02/03/enterprise-vs-consumer-g…

suggests that the issue is write cycles, but I would like clarification on that. The other suggestion relates to persistence in the event of power failure, but that seems to me to be a property of “enterprise” overall hardware vs the flash itself. The article suggests that primarily read-only applications like web servers would be a good use case for consumer grade. There is also a question of the failure mode … if a module were able to announce that it had had all the writes it could take and one could replace the module in the overall device relatively cheaply, then consumer grade might still be a good alternative … but I am a little dubious about knowing that a module was going off.

If Jeppesen is distributing using flash the data is pretty static. Failure of consumer flash is over many repeated write cycles. So I wouldn’t give all that much weight to what Jeppesen does.

I would think that with various RAID techniques the flash reliability concerns are likely minimized. Add it a little of self-monitoring and log the error rates and you likely can predict when a failure will be significant enough, and then have the device call the mothership and dispatch a technician.

That may be an oversimplification, but it seems like it would be doable with current tech. Heck this was doable 30 years ago when IBM and Amdahl used to do it with their systems.

-----
Invest wisely my friends