Hey, I was reading these comments, and started thinking. Couldn't the effects of wear and tear be reduced by implementing some sort of data redundancy or error correction? CD's do it, DVD's do it better, and similar techniques are being used in RAID today. Sure, some cells would lose their data, but we'd be able to recover that data, possibly transparently to the user. The downside is that you get a slight reduction in storage space. What do you guys think?

freakin awesome!

This has been possible for years. To make it usable on a wide range of equipment, the price has to come down. Flash is now at about $50/GB... if this drive costs $800, it's still not interesting.

Anyone out there old enough to remember hard cards? They were memory-based hard drives. POS things that didn't last because they were unreliable among other things.

"Nothing for you to see here. Move along."

Back in 1986, a hard card was an actual hard drive mounted on an ISA card for ease of installation, back when hard drives required jumpers, manual setting of drives/sectors, etc. The hard card did it all for you, but at the heart of it was still a hard drive. It was pretty reliable, but very expensive (about $750 for 20MB)!

I used to have a 40mb MFM drive that was embedded in an ISA controller card. It was an interesting critter. I do remember the cards you are referring to, and actually I believe that Quantum makes solid state disks today.

Does anyone remember the 1MB sram "floppy drive" project from a magazine back in the late 80's?

:-)

Yeah, that. I had one of those, they were very expensive.

http://www.microstar.net/museum/hardcard50.html

That's the model that I had.

ANY COMPANY MAKING SOLID STATE MEMORY/DRIVES NOT BASED ON FLASH, which has limited write cycles ???

The Plus Hard Cards were ISA form factor full slot cards with a physical (MFM) HDD, not a memory based storage medium. I never found them to be less reliable than most of the MFM half height drives such as the ST-225 series. There were some attempts at bubble memory hard card sorts of devices, but they never caught on in any big way, and improvements in easier to manufacture and more reliable non volatile technologies pretty much left the whole approach in the gutter. Neat stuff to play with at the time though :-)

Stan

What about the fact that you can only erase and rewrite to a flash drive so many times before you can no longer write to it? Unless either they've done something about it or that threat is overblown, I don't think this could totally replace a normal hard drive.

What limit have you heard about on flash drives? I've worked in IT as a network engineer for 10+ years and NEVER heard this. Illuminate me...;-) I mean eveything has a MTBF spec.

MTBF specs vary based on the manufacturer and the calculation used; the following are some sample specs I have found:

Pretec --> MTBF 500,000 hours (powered on)
Simpletech --> MTBF 1M+ hours
SanDisk --> MTBF 3M+ hours
BitMicro --> MTBF 2M+ hours

But the most reliable and experienced guys around are BitMicro, and this is what they pubish for one of their flash drives:

http://www.bitmicro.com/...urces_flash_ssd_db2.php

Example #2: Write Frequency in MB/sec

E-Disk® PB Size = 16 KB

I/O Block Size = 64 KB

Write Frequency = 6,016,204,800 KB per day (68 MB/sec)

E-Disk® capacity = 155,648 MB
Number of Flash chips = 608
Size of Flash chips = 2048 Mbit or 256 MB or 262,144 KB

Number of writes to Flash chip = 64 KB / 16 KB = 4
Total E-Disk® physical blocks = (262,114 / 16) x 608 = 9,961,472
Total max writes to E-Disk® drive = 9,961,472 x 1,000,000 = 9,961,472,000,000

Endurance (in days) = 9,961,472,000,000 / (4 X (6,016,204,800/64)) = 26,492 days
Endurance (in years) = 199,229 days / 365 = 72.59 years

It is interesting that SamSung has developed something very similar to what BitMicro has had for some time.

The gentleman with the first post is correct, flash has a limited number of read/write cycles.

The number used to be about 1000 cycles, it may have gone up since, but the limitation is still there.

Network Engineer for 18 years.

To clarify any confusion caused by all the MBTF stats quoted here.

The limited numbers of write cycles of Flash memory has absolutely nothing to do with MBTF.

You can reach the limit of write cycles in one day if you write to the flash often enough, or you can reach it in 10 years if you write infrequently.

I agree. The MTBF (Mean Time Between Failures) is a theoretical calculation derived from the reliability metrics of the individual components, their respective failure modes, at some target operating temperature.

The max write cycles is an absolute limit usually based on the semiconductor properties within the memory device.

I did not mean to imply that these two were one and the same.

Please note that BitMicro stated a write frequency of 6,016,204,800 KB per day (68 MB/sec) - this is continuous usage of the device at a very high data rate. The endurance is much more than the lifetime of any computer system --> (in years) = 199,229 days / 365 = 72.59 years

I believe the gentleman that posted the MTBF statistics was trying to show you, even with high usage, they're still usable after such stressful testing. If they're writing as much as the stats say, they've far exceed the 1000 cycles number mentioned before.

I would be willing to bet big that the drive has some form of wear levelling.

http://en.wikipedia.org/wiki/Wear_levelling

If the filesystem and/or the flash drive's firmware is designed carefully, and distributes writes evenly across all of the flash disk's blocks, then its life can be extended enough that it is for all practical purposes infinite.

However, if one were to treat a flash drive as a normal hard drive, frequently overwriting certain blocks that had often-modified files, one could "burn out" parts of the disk long before the manufacturer's specs.

I'm guessing modern flash drives dynamically remap physical blocks to logical blocks based on frequency of writes to the logical block. (Where the drives store THIS information is another matter entirely ... perhaps they have an onboard SRAM and battery for this purpose.)

I recently had to look that up for some applications. # of write cycles modern flash cards are good for usually vary between 10,000 and 1,000,000 depending on manufacturer and model. Note: many applications tend to re-write multiple time when saving new information to a disk, burning though those cycles faster than might be obvious.

WHAT IS A FLASH DRIVE

I hadn't heard of this either, but an entry at wikipedia seems to support the statement.

"When compared to a hard disk drive, a further limitation is the fact that flash memory has a finite number of erase-write cycles, so that care has to be taken when moving hard-drive based applications, such as operating systems, to flash-memory based devices such as CompactFlash."

http://en.wikipedia.org/wiki/Flash_memory

You're missing the point - data write rate in the sense of how much data you pump through it from an application level isn't the problem.

The problem is that in a "normal" filesystem, certain blocks are re-written every time you read the file. Remember, directories are files too, so even just listing a directory modifies those blocks. In this way you could easily end up with a flash system that's reached the end of its write cycle very quickly.

One culprit for which there is a workaround is filesystem access-time updates. The workaround in this time (on linux) is to mount the filesystem using the 'noatime' option to the mount command.

As someone else posted, wear-leveling is another (more general) solution, if the hardware provides it.

LOL!!! I'm sorry but it look comical with this long thread talking about flash drives and then on the end is:
"WHAT IS A FLASH DRIVE" hehehe.

To answer you question I got this qote right out of www.WikiPedia.com

"A flash drive is a solid state read and write device that attaches to a computer and is usable as a standard hard drive. These devices come in many forms and commonly attach via IDE or USB. A specific type of flash drive is a keydrive which is small and easy to plug in to your computer."

Anotherwords, its solid state because it has no moving parts. Unlike your normal drive has a spinning platter and an access head going back and forth much like a needle on a vynil record.

Check out http://en.wikipedia.org/...lash_memory#Limitations

Haven't they gotten over a billion now?

You have to understand that the internal memory controller of the hard drive takes into account wear leveling - we are not talking about raw flash memory cells, but about an integrated hard drive which uses flash memory.

Wear Leveling extends the lifetime of NAND Flash devices because it ensures that even if an application writes to the same virtual blocks over and over again, the program/erase cycles will be distributed evenly over the NAND Flash memory.

With wear leveling, you can achieve the stated lifetime of 50+ years. The MTBF calculation and lifetime statistic has to take into account the failure modes of the system as a whole, including the wear leveling technology that extends the lifetime of the flash memory cells.