Just read this basically puff piece on a new company for software for multi-SSD hard drives.
I know it is a journalist after all, but the developer interviewed is a big shot. SSD development is enourmous in Israel.
Just read this basically puff piece on a new company for software for multi-SSD hard drives.
I know it is a journalist after all, but the developer interviewed is a big shot. SSD development is enourmous in Israel.
ETA: http://www.timesofisrael.com/how-an-israeli-start-up-beats-flashs-dirty-little-secret/
Yes, Flash has both a maximum data retention time, and a maximum number of write cycles before the data retention time becomes unacceptably short. This has been true for all solid-state non-volatile memory.
It’s real, but not much of a concern for ordinary desktop use. Basically each memory cell can only be overwritten so many times before it wears out. This limit will depend on the exact type of flash memory used. Consumer drives are typically rated for 1,000 or 3,000 writes. That’s not so bad actually, since (1) the SSD uses “wear leveling” to make sure that no part of the drive wears out before the rest, (2) even if you were constantly writing data as fast as possible, it would take months or years to wear out the drive, and (3) under ordinary desktop scenarios, it’ll take a decade or more to wear out the drive.
For certain types of server use, on the other hand, the finite life span of a SSD is a real problem.
All storage devices experience “fade”, including flash memory and hard drives (this is often the cause of “bad sectors” in old failing drives which otherwise have no physical damage). As for whether it is a concern, it is if you are planning to use it to archive data for many decades but otherwise it shouldn’t be a problem if you keep the drive cool; as for why I say this, here is a paper (PDF) that shows that there is a very steep logarithmic relationship between temperature and retention time (Arrhenius equation, usually said to double the rate of a reaction or halve lifetime for every 10°C increase but the interval can be more or less, as with flash memory; this also enables reliable retention time estimates without spending years testing them). Thus, if the device is specified to retain data for 10 years at 85°C, at room temperature that pretty much guarantees centuries of data retention. In particular, note the difference between 25 and 55°C:
So a reduction (increase) in temperature of just 30°C increases (reduces) lifetime by a factor of 19-fold (roughly a factor of 2.675 per 10°C, thus 10 years at 85°C becomes about 3,600 years at 25°C).
Of course, if you write frequently to the drive it will reduce the retention time until it fails (the good thing though is that for frequently written data a write failure occurs on write, so the drive will detect this and reallocate the sector).
My OP has fade in quotes, and yours doesn’t. 
I’m taking it that temperature will effect the magnetic strength?
But if temp is controlled, what causes the fading?
Thermal energy.
Unless the device is at Absolute Zero, the atoms still have thermal energy, and this causes magnetic devices to loose their polarization, and electronic devices to lose their trapped charge.
What happens is that each time a cell is drained, a very very tiny bit (relative to fully charged state) gets left behind. Over thousands of cycles, this charge builds up (kind of like dust on top of a cabinet) until it can no longer detect the difference between charged and not-charged state.
There are tricks SSD drives use to extend the life of cells, including cell rotation (insures cells are written to equally), overprovisioning (a 500GB drive might actually have 600GB of cells), and smart-caching (collects write operations until they’d fill a whole block (the smallest individually addressable unit). This is all done by the software of the drive itself and is invisible to the operating system (and therefore any hardware).