Slashdot Mirror
Ultra Fast Disk Drives With No Moving Parts
saccade.com writes "Let's face it, the slowest part
of PC's today is the disk drive. Bit
Micro has come up with a nifty solution - flash memory based
disk drives available in typical
disk
form-factors. These e-disks are electrically compatible
with ATA, SCSI, etc. but run orders of magnitude faster - access
times down to 40 usec and transfer rates over 100 MB/sec. What's
the catch? Cost. Currently going for just under $1K/G, a 30G model
I recently held in my hand was worth much more than my car. However,
as flash memory prices drop, so do the price of these drives.
Within the decade the spinning hard disk may go the way
of the floppy and CRT."
This isn't exactly new. They've come down substantially in price and gone up in volume, but these have been around for years. It is my understanding that the most significant use was (is?) laptop drives for extremely rugged, shock-resistant portables.
That what was all this school was for... to teach us how to solve our own problems. -- janeowit
Because I'm pretty sure most of us were aware of high cost flash media disks.
I wonder how long you can beat at a device like this in a server environment before it croaks. I'd give it no more than a year life expectancy, but hey, I'm feeling pessimistic.
I am feeling fat and sassy
I need an EE to build an ata interface to a raid series of about 100 flash either (SD or compact). Now allow the end user to plug in how many cards he wishes and just use them. Imaging that if you have a raid 5 setup of say 128 256mb cards costing about $40 each would cost about $5000 1/6th of the $30k and it is end user upgraded and so cool to be able to ad more storage instead of rebuilding a whole computer and drive.
SSD (Solid State Disk) has been around for over 30 years. Every so often it is billed as the "spinning-rust"-killer which has yet to be borne out. It's a great idea but so far rotating media has managed to improve enough to make SSD uneconomical.
Within the decade the spinning hard disk may go the way of the floppy and CRT
As an aside, my CRT is still firmly wedded to my desktop, and won't budge until flat screen technology has caught up. It's come a long way, and may be good enough for less demanding applications, but it's got a way to go before I have a flat screen on my desk...
"The invisible and the non-existent look very much alike." -- Delos B. McKown
...to store data by etching bits with a stylus into Faberge Eggs.
Are we done yet with the whole 'floppies are dead' stories? I regularly use floppies because it's easier to plop in a floppy, copy one file and pop out the floppy than it is to put in a USB drive, wait for your pc to recognize it (don't know about Macs), copy the file then have to correctly disconnect the USB drive
What about those machines which don't have USB drives or who aren't on a network? What then? Floppies will be around much longer than anyone thinks and for good reason.
We will bankrupt ourselves in the vain search for absolute security. -- Dwight D. Eisenhower
And here I thought you had to pay to run an ad on Slashdot...
I think there is a world market for maybe five personal web logs.
The slowest part of PC's today is the disk drive.
No, the slowest part of PC's today is the user interface. The rate at which a user enters data (via keyboard/mouse) is a fraction of the rate at which a user thinks. (Your mileage may vary, of course.)
-kgj
-kgj
Within the decade the spinning hard disk may go the way of the floppy and CRT.
You mean it'll still be the default option on most new PCs and in use by ~90% of PC users?
100,000 writes isn't gonna last long in todays bandwidth intensive video/mp3 world
no moving parts and non-magnetic media is a worthy goal but until we can cure terrible storage lifetimes they wont be much use if i have to worry about the mess backups of backups, as we know from sci-fi all it takes is a big EM burst from the sun and everything you and i have done is gone !
future generations will look back at us and say "they used to store it on WHAT !?"
The reason hard disks etc are seperate devices is because they have mechanical parts that require motors etc to work. If this is going to be replaced by memory chips then why not just integrate the whole lot on the motherboard as just another plug in memory module? Why make it slower by passing it through SCSI or ATA not to mention the extra cost of including the interface electronics?
I've always found the best way to deal with the problem of slow disks is to max out the memory in the PC and use a hefty chunk of it as a RAM disk. When done or needing to backup, tarball the whole disk, write it once to the hard drive.
:)
Of course, this assumes you're working on a stable OS with decent tools and good memory management. If you're not, you can be.
Everything in the Universe sucks: It's the law!
This "disk drives will be obsolete" assumes that disk drive prices are flat. Drive prices are one of the few things that has (if anything) beaten Moore's Law. Eventually they'll probably flatten out - but not yet. The "death knell of rotating media" has been sounded more times than I can remember. Anyone remember the front-page stories that by late 80's bubble memory would have replaced hard disks? :-)
I know that 10000 writes seems like a lot, and perhaps it is. Anyone knows how this figure looks for normal harddrives?
Still it seems to me that the limited number of writes sets the biggest limitation.
What about those machines which don't have USB drives or who aren't on a network? What then? Floppies will be around much longer than anyone thinks and for good reason.
What about those machines which don't have floppies?
Seriously, I haven't put a floppy into a machine in the last 6 years. They're totally unnecessary nowadays. They're about useless for transporting documents for the simple reason that the majority of useful documents exceed the size of the floppy nowadays.
And USB drives are much cooler than you seem to make them out to be. Plug the thing into the USB connector in the front, it mounts, you copy, you unplug the thing. Yes, you might have to wait a second or two for it to recognize and mount the thing, but that's better than waiting for at least 90 seconds to copy 1.4 meg to the slow-as-hell floppy.
Floppies once had limited usefulness as being the only easy way to bootstrap the system. Boot from the floppy, format the hard drive, install the OS. Now that every mobo can do CD booting, I no longer need boot floppies, as I can have boot CD's instead.
Within the decade the spinning hard disk may be capable of holding terabytes, or even petabytes, on a single platter. And it will be orders of magnitude cheaper than solid state storage as we know it. I doubt that hard drives will go the way of the dodo anytime soon.
Just as a comparison, look at how many backup solutions still use tape media (and use it very effectively and cheaply, I might add).
Have fun: Join D.N.A. (National Dyslexics Association)
...printer.
Technically, a printer is a peripheral, not a part. Whatever. All printers are evil: Too slow, too big, too expensive, too quirky. Ackk.
-- Slashdot: When Public Access TV Says "No"
Sure, hard drives are slow, but that's not my main problem with them. They *are* a bottleneck, but since most applications get the hard disk access "out of the way" at the very beginning and load everything they need into RAM, I could deal with slow hard drive technology for the rest of the forseeable future, if only...
... they were reliable. Hard drives are the only PC components that have ever died on me. Actually, that's not quite true - I had a CD-rom die once, and a few fans here and there; what do all these have in common? Mechanical parts. And when it comes down to it, what do most users value most in their computers? The files on their hard drives. Spinning death traps is what they are. Spinning, clicking, grinding death traps.
I don't know much about flash memory technology or the reliability associated with it. I don't give a hoot how fast it is. If it's solid state (no moving parts) and can guarantee me it won't one day decide to utterly destroy itself, I'm sold.
First of all, the technology used in a product like this is not radically different from existing flash solutions. The big problems are cost and limited use -- flash memory (transistors with high voltage-forced states) can only be toggled a limited number of times. So there is a limited number of write cycles for the faster types of non-volatile solid state memories.
That problem can be reduced by padding devices with large amounts of RAM (write caching). But the breakthrough is coming soon, with new flash technologies that are better designed for continual writes (without compromising speed). From what I've read in IEEE Spectrum, the better technologies suited for mass storage are in research labs right now, meaning maybe 5 or 10 years til market.
I dug a bit and found this in the manufacturer's FAQs:
QUESTION: What is the lifespan of the E-Disk® flash drive if wear-leveling algorithm is not utilized? How much improvement will BiTMICRO's wear-leveling algorithms make to this number?
ANSWER:
The wear-out life of an E-Disk® flash drive is directly proportional to the number of flash memory physical blocks in the device. The greater the number of flash memory blocks in the flash drive (and therefore total capacity), the longer the wear-out life of the device. As an example, arithmetic computation will show that a 34GB E-Disk flash drive fitted with flash chips rated at an endurance limit of 1 million erase/write cycles will have an endurance life of 1,024,000,000 seconds (or 32.47 years) when written continuously at 34MB/sec (or 2,937.6GB Erase/Write per day). This is the worst possible scenario where all I/O is 100% write and caching is disabled. E-Disk erase/write endurance can be more than 15 times the computed value if the multiplier effects of full associative caching and the results of BiTMICRO's accelerated erase/write endurance verification and testing are included.
Old enough, so the first 'generation' of SSD companies is already out of business. E.g. Platypus (I think that was the name) build RAM based solid state drives, some of them in the right shape and with appropriate disk interfaces to match existing disk drives.
I looked into SSD for a database at one point. But I found that you can get almost the same performance by using lots of drives in a fast RAID setup. Striping the content over multiple disks does wonders! And its much cheaper.
E.g. look at something like a 12 disk setup with RAID 5+1. You got a full mirror, and essentialy 4-8 times the speed of a single drive. So you are already close to the 'order of magnitude' they SSD drives claim.
---- join dshield.org Distributed Intrusion Detec
take a look at this raid 0 floppy setup: http://ohlssonvox.8k.com/fdd_raid.htm
yes, I know that it would cost more and we would still have moving parts. It's also slower.
But just imagine a room with ~21300 FDD (30 gigs) stacked to the ceiling blinking and spinning like mad.
I don't read replies by ACs.
According to their web site, their "Patented Wear Leveling" algorithms attempt to spread write operations over the disk. My guess if you have a frequently written file/record/whatever is that it doesn't write it to the same place each time. It also looks like they have a "Flash Wear-Out Monitor" to warn you when the device has exceeded 95% of it's MTBF rating, though they say that the device may last beyond the rating. Also, looks like their "Automatic Bad Block Remapping" moves data to spare blocks if a block fails. So, yeah, like you said, they work around the dead bits remapping them to a new area, as well as constantly spreading write cycles across the device. Looks like they've really thought this through. Of course, so long as the price exceeds that of spinning platters, it'll be a niche product.
As far as "Within the decade the spinning hard disk may go the way of the floppy and CRT" goes, I guess that means that there will be other/better/different choices than spinning platters, but they'll still be more expensive and spinning platters will still be the norm. Looking forward to the status quo, I guess!
"Single-level cell" flash memories can manage over 100,000 writes per sector. "Multi-level cell" flash memories, which slightly lead single-level on the density/time curve, can manage only roughly 10,000 writes per sector. Learn more about the difference between single- and multi-level cell flash memory.
With this thing rated at up to 25,000 IOPS, is would seem that they might not last all that long (4 seconds?).
Yeah, with tens of thousands of writes to the same sector. CF flash memories already perform some sort of wear leveling to spread repeated writes over multiple sectors. Yeah, it's more difficult for swap files, but I expect that rather than use a swap file on flash memory, PCs with solid-state storage will use more volatile memory.
Maybe that's why Activision won't sell me a version of Doom 3 on 1,300 floppies. Why didn't anyone tell me this before?
Probably, a better HD-replacement solution would be based on MRAM, which is being steadily developed and is going to become available quite soon (the article linked mentions late 2004).
Computers make very fast, very accurate mistakes
Compact Flash is already IDE. It's just that the pinouts are different. You can buy an adapter for ~ $20. The previous poster was correct about the maximum number of writes though. I have a system that I use compact flash to boot off of in RO mode. My system boots fast and I don't write to the disk.
Encryption: I may not agree with what you say, but I will defend your right to encrypt it...
all these elements are "flat", that is they are one structure deep. This new tech coming up, if someone can perfect it, uses multiple layers to make the flash array several layers deep. Thus you could (in theory) shrink your die size while increasing the memory density.
This turns out not to help much. Multi-layer chips add mask steps roughly in proportion to the number of layers. While you save on the cost of wafer area, your processing steps cost a lot of money too, so you rapidly reach a point of diminishing returns. Building multi-layer devices also requires making transistors on epitaxial silicon layers, which generally have far worse performance properties than the monocrystalline wafer (even SOI processes generally work by building devices on a silicon wafer, and either flipping the chip and back-etching or using a buried oxide layer, as opposed to depositing a silicon film).
3D chips have been a holy grail for density reasons for decades, but they turn out to be expensive to manufacture and poorly-performing for the reasons noted above, and for microprocessors, at least, they're now a pretty much obsolete solution, as heat generation is what limits chip performance (and a multi-layer chip gives you that much more heat generation per unit area).
If your company can pull it off in a useful way for storage, they'll deserve kudos, of course.