NAND Based Flash RAID

  • Thread starter Thread starter dan555smith
  • Start date Start date
nickalex said:
I suggest thast you all go and have a look at panasonics P2 PCMCIA
cards used in there pro video cameras utilizing 4 SD memory cards and
a RAID 0 controller getting speeds in excess of 600Mb/s
Click to expand...

That's still not faster than a HD can do.
 
Yeah, after I checked more into this (flash/pen/key drives), it really
isn't any kind of faster. not only isn't it faster, it is so much
slower that it isn't even a mediocre option. 20-30 MB a second is about
as good as it gets. if the access time to the device is 400MB a second,
but it can't transfer or read/write at better than the 20-30MB/sec it
really is a dumb idea. When I asked, I had not read THG's flash memory
guide. or any others. I wanted something low power (not another hard
drive) but I guess my Idea is too soon. Nickalex has a great Idea, I
will check into that...I saw another thing some time ago on THG which
was a 6 pack of laptop drives in 1 5.25 slot as a raid setup. Novel
Idea, but pretty expensive too. again, it takes away from the power.
then there is the 1-8 GB of ram on a pci card that gigabyte came out
with. sounds good, but haven't gone over the specs. maybe they will
have a PCI-e x8 card...anyway...enough dreaming for now.

By the way, it wasn't for gaming, it was for applications on a desktop,
or a server...as an example.
 
John said:
Hello, Dan:

If you're considering flash memory cards (used by digital cameras, for
example) or USB "key drives," forget them. They both have a relatively
limited number of rewrite cycles -- and hence, would be quickly worn
out, by a page file.

Further, they're somewhat slower than HDD's, despite being "solid
state." (The tiny IBM/Hitachi "Micro Drive" avoids these issues, as it's
an actual electro-mechanical device, just as a reqular hard disk is.)

A real SSD (solid state drive) features hardware similar to system RAM,
instead of flash chips. It is, therefore, much faster than any
conventional HDD...and correspondingly, far more expensive, as well! <g>

Cordially,
John Turco <[email protected]>
Click to expand...


Slight correction: SSD refers to "solid state disk," more commonly.


Cordially,
John Turco <[email protected]>
 
Ok, lemme help:

1) Flash is not faster than a physical hard drive is capable of. If
you want a solid state storage solution, you want a RAM-based
IDE/SATA drive. They get very close to the theoretical limits of
each protocol (being memory, reads, writes and accesses take
something on the order of nanoseconds)
2) Rather than adding a 2Gb paging file, you could try bumping your
machine's RAM up. A good reason for this is that no matter HOW big a
paging file or HOW fast the protocols between the processor and the
pager, RAM is faster. RAM doesn't have abstraction layers for the
data to punch through, and it sits logically closer to your processor
(the leads are shorter than those to a drive.
3) If you're looking for flash-based IDE devices, they are floating
all over the interweb.
4) Compact flash is actually 44-pin IDE with extended protocol
support.
5) MMC/SD will do a kind-of raid-0 all on their lonesome - just solder
together a few leads; there's instructions on how to do so on the
inter net.
6) Learn to use google. Google knows more than you do.
 
Fordiman said:
Ok, lemme help:
Click to expand...

You think?
1) Flash is not faster than a physical hard drive is capable of. If
you want a solid state storage solution, you want a RAM-based
IDE/SATA drive. They get very close to the theoretical limits of
each protocol (being memory, reads, writes and accesses take
something on the order of nanoseconds)
2) Rather than adding a 2Gb paging file, you could try bumping your
machine's RAM up. A good reason for this is that no matter HOW big a
paging file or HOW fast the protocols between the processor and the
pager, RAM is faster. RAM doesn't have abstraction layers for the
data to punch through, and it sits logically closer to your processor
(the leads are shorter than those to a drive.
Click to expand...
3) If you're looking for flash-based IDE devices, they are floating
all over the interweb.
Click to expand...
4) Compact flash is actually 44-pin IDE with extended protocol
support.
Click to expand...

Uh no. It's neither 44-pin nor is the pin-out identical to IDE.
You need an adapter.
 
i know this concept has no real world function right now, but once HD
are made without moving parts there will be less room for erro
(especially when using them in turbulent situations). so i wondere
the same thing. HDs with no read heads can't crash as easily, an
can be more reliable in mobile storage units. . . and right now flas
is only at a gig so i wondered if you could raid them to make them
competent HD that could match the 5400 rpm hds that are standard i
laptops. i really think that the future is no moving parts--i
leaves less room for error. i know someone will call me a dumbass
but **** it
 
jbm666 said:
i know this concept has no real world function right now, but once HDs
are made without moving parts there will be less room for error
(especially when using them in turbulent situations). so i wondered
the same thing. HDs with no read heads can't crash as easily, and
can be more reliable in mobile storage units. . . and right now flash
is only at a gig so i wondered if you could raid them to make them a
competent HD that could match the 5400 rpm hds that are standard in
laptops. i really think that the future is no moving parts--it
leaves less room for error. i know someone will call me a dumbass,
but **** it.
Click to expand...


Simple.

Look at the specifications for write speed and the need to erase a block
before writing it. Look at the time to erase a block.

Look at the block size and any requirements for sequential access in the
block.

Look at the number of write cycles that are supported.

Look at the cost of 200 GB of FLASH compared to 200 GB of hard disk.

Then make a proposal.

craigm
 
craigm said:
Simple.

Look at the specifications for write speed and the need to erase a block
before writing it. Look at the time to erase a block.

Look at the block size and any requirements for sequential access in the
block.

Look at the number of write cycles that are supported.

Look at the cost of 200 GB of FLASH compared to 200 GB of hard disk.

Then make a proposal.

craigm
Click to expand...


Hello, Craig:

Finally, take a look at flash memory's decidedly low number of rewrite
cycles, in comparison to actual hard disk drives. You'll see that it's
only suitable for storage purposes, whereas a true SSD (solid state
disk) is a RAM-based device -- and will, therefore, vastly outperform a
HDD.


Cordially,
John Turco <[email protected]>
 
Eric said:
Absolutely clueless.
Any solid-state HD uses ECC and remapping, just like magnetic drives.
If it is flash-based, it also uses wear-leveling: http://www.lvr.com/mass_storage.htm
Flash drives have taken over mobile drives in extreme environments.
Click to expand...


Hello, Eric:

Well, here's a quote from the Web page you just cited:

"Wear Leveling

Flash memory is a popular choice for smaller drives. Flash memory can
withstand a limited number of erase/write cycles, however. A typical
guaranteed number of cycles is 100,000. Firmware that repeatedly writes
to the same addresses will quickly wear out that portion of the memory.
To extend the life of a Flash-memory chip, firmware can implement wear
leveling, which uses various techniques to access all areas of the
memory equally."

Nowhere does it imply that this "life extension" can allow a flash-based
"disk" to replace a HDD, in a computer...which, for example, would
access it far more often than a digital camera.

No problem, though, for a (considerably more expensive) RAM-based SSD.


Cordially,
 
John Turco said:
Hello, Eric:

Well, here's a quote from the Web page you just cited:

"Wear Leveling

Flash memory is a popular choice for smaller drives. Flash memory can
withstand a limited number of erase/write cycles, however. A typical
guaranteed number of cycles is 100,000. Firmware that repeatedly writes
to the same addresses will quickly wear out that portion of the memory.
To extend the life of a Flash-memory chip, firmware can implement wear
leveling, which uses various techniques to access all areas of the
memory equally."

Nowhere does it imply that this "life extension" can allow a flash-based
"disk" to replace a HDD, in a computer...which, for example, would
access it far more often than a digital camera.
Click to expand...
My 4 year old SCSI drive has 3 billion writes. It has 30+ million sectors. Do the math.
 
Eric said:
My 4 year old SCSI drive has 3 billion writes. It has 30+ million sectors. Do the math.
Click to expand...


However, that would give you an average, not the highest number of
writes that any one block may have seen. Some sectors were probably
written once. Others will have been written much more. (Think about the
directories and file system areas.)

To do wear levelling, the controller need to keep a map of every block
written. How big do you think that is and where would you store it?
Where would you put this feature? In the host software or the 'drive'?

Do you realize that FLASH needs to be erased before it is written? That
also slows things down.

There are reasons why nobody does this. Ever wonder why?

craigm
 
However, that would give you an average,
Click to expand...

Exactly.
What you expect the number of times every cell to be
written to if wear leveling is doing what it claims to do.
not the highest number of writes that any one block may have seen.
Click to expand...

Doesn't matter.
Some sectors were probably written once. Others will have been written much more.
(Think about the directories and file system areas.)
Click to expand...

That's not what he mentioned those numbers for.
To do wear levelling, the controller need to keep a map of every block written.
Click to expand...

And a lot more too. It also needs to know where that block is currently.
That's why they have "Dynamic mapping of any logical sector to any physical sector".
How big do you think that is and where would you store it?
Click to expand...

Are you questioning wear leveling?
Where would you put this feature? In the host software or the 'drive'?
Click to expand...

Drive obviously.
Do you realize that FLASH needs to be erased before it is written?
Click to expand...

What has that got to do with wear leveling.
That also slows things down.
Click to expand...

Not if you use up all the free blocks (pre-initialized, ie erased) first
(data unit rotation?) and erase the previously abandoned blocks in
times that the memory is not being addressed (garbage collection?).
There are reasons why nobody does this.
Click to expand...

So you also question John's quote
"Flash drives have taken over mobile drives in extreme environments" ?
 
Folkert said:
What has that got to do with wear leveling.


Not if you use up all the free blocks (pre-initialized, ie erased) first
(data unit rotation?) and erase the previously abandoned blocks in
times that the memory is not being addressed (garbage collection?).


So you also question John's quote
"Flash drives have taken over mobile drives in extreme environments" ?
Click to expand...


Hello, Folkert:

Eric Gisin wrote that comment about "mobile drives in extreme
environments," not yours truly. Nonetheless, as my reply did quote
his message, you may still be technically correct. <g>


Cordially,
John Turco <[email protected]>
 
Back
Top