Understanding the effects of power failure on flash-based SSDs

[Tom Del Rosso]

No one is disputing that capacitor-backed RAM would also be considered
NV, but flash is also NV. Why make things extra convoluted by having
one type of NV memory backup another type? Isn't it better just have
that capacitor keep up enough charge so that the flash memory itself
so that it can complete?

Do you mean so it can complete its write cycle? That's part of the benefit
of what I suggested.

But the two main benefits, which I forgot to mention, are that it would
extend the life of the flash, and make writes faster. It's a cache as well
as a buffer.

The fact that power-loss causes the loss of previously written data is yet
another reason. You could pull the plug in the middle of a write cycle, and
complete that cycle without corrupting older data.

There is another type of NV memory called MRAM which is apparently the
fastest writer of all, nearly as fast as SRAM for reading and writing
operations, but it has some problems in being miniaturized, so far.

Oh yeah, magnetic memory keeps popping up again in different forms. I
remember when bubble memory was going to be big. Some computers were built
with it. I think MRAM would have distinct cells like other RAM, while
bubble memory had amorphous magnetic regions on a chip.

 

[Yousuf Khan]

Do you mean so it can complete its write cycle? That's part of the benefit
of what I suggested.

But the two main benefits, which I forgot to mention, are that it would
extend the life of the flash, and make writes faster. It's a cache as well
as a buffer.

The fact that power-loss causes the loss of previously written data is yet
another reason. You could pull the plug in the middle of a write cycle, and
complete that cycle without corrupting older data.

Well, that is what happens in some forms of SSD's, they have a small RAM
cache that buffers writes to the flash memory. In other SSD's, namely
the Sandforce ones, all of the performance is achieved by interleaving
the flash ram chips with each other heavily, avoiding the need for a big
RAM cache.

Yousuf Khan

 

[Arno]

It might be due to the fact that the other members of the RAID set are
on such a different plain of performance than the SSD. The SSD might be
dropping out because the RAID software/firmware finds it finishes much
faster than the HDD, and that would look like an error occurred on the
drive, according to the RAID software.

Not at all. It is not the RAID layer that fails. It is the plain
SATA interface. And I have to say your theory is something
pulled out of thin air...

Arno

 

[Arno]

Well, that is what happens in some forms of SSD's, they have a small RAM
cache that buffers writes to the flash memory. In other SSD's, namely
the Sandforce ones, all of the performance is achieved by interleaving
the flash ram chips with each other heavily, avoiding the need for a big
RAM cache.

Aehm, Sandforce has even larger buffers, as they do data compression.
Interleaving is something many other controllers can do.

Arno

 

[Mike Tomlinson]

There is another type of NV memory called MRAM which is apparently the
fastest writer of all,

Do you mean memristors?

 

[Yousuf Khan]

Do you mean memristors?

Yes, kinda, they fall into that category:

Magnetoresistive random-access memory - Wikipedia, the free encyclopedia
http://en.wikipedia.org/wiki/Magnetoresistive_random-access_memory

Yousuf Khan