SSDs Have Bleak Future, Says Researchers

Page 6 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.



The problemsthey are talking about aren't due to just needing to improve the item. The problem is due to leakage at the atomic level. You cant go smaller than 5nm because the electricity leaks and shorts out the chip in question. There is NO work around for this issue and its whats driving the search for new materials to make chips out of. Graphene is one material being looked at. If you made an i7 2600k out of graphene you would be able to push the chip to 100Ghz but thats the top end, you still couldn't make the die size any smaller because of the same leakage.

Now HP and a few other chip makers are looking at photon based CPUs because you dont have the same leakage problem and the chips would be many times more powerful than similar electricity based CPUs.
 
The term SSD could be applied to any solid-state memory based storage device, not just NAND flash based ones. It seems shortsighted to claim the doom of SSDs based on the short comings of one memory type. What about PCM, ReRAM and STT-RAM based SSD's?
Click to expand...



They all suffer from the same problem at the atomic scale. You can only go so small before you get leakage and the chip shorts out. There is NO work around using current technology
 
It never fails that someone will predict a limit on some technology evolution and a few years later it's proven wrong. There has often been a trade off for performance or capacity when it comes to any storage technology. Right now that trade off is not high enough to drive anyone away from the SSD drives coming to the market and when that happens innovations will be necessary to push the importance of performance gains.

Right now the focus of the industry is more capacity for less cost so this discovery makes perfect sense in the current market. As the SSD drives begin to compete more with each other and less with platter based drives things should begin to heat up on the performance situation.
 


You're correct, but you neglect the fact that the different technologies scale differently. NAND doesn't scale as far as many other technologies, especially the upcoming ones. The memories that don't store data with electricity have incredible promise in this regard because they don't worry about leakage as much.

It's also worth saying that leakage is not the only problem with scaling to smaller and smaller processes. One such problem is the difficulty of detecting the smaller and smaller electrical charges. This is another reason why the memories that don't store data with electricity have great promise.
 
This guy's crazy if he thinks we're going to be using the same material 12 years from now.
To be honest, I don't think we'll be using SSD's in the future as they are now, at all. I imagine there will be relation storage methods for describing an item in order to massively compress data. Of course, we already use that now by just having file types, but it's extremely inefficient and far from uniform. But to incorporate it would require processing inside the HDD, which we don't have.

But it's just fun to think about. 😛
 
I expect them to solve the latency similar to the procesor cache latency problem: which is Leveling. So once latancy becomes getting an issue we will see Leveled SDD where write are first done in a low latancy SDD and transfered by the onboard memory controller when more space is needed.
 
All blah blah blah blah to me this is from someone looking for a feather in her hat to graduate college. I guess she never had one of those record players (HDD) up and die just because it wanted too or decide oh I have a multi sector error on your 3GB PST file and its gone. Maybe drop her cup of Starbeaters on the sweet spot on a laptop and watch the doom ensue. This one is my favorite as well, the 2.5inch USB portibles get dropped into a bag, and boom its dead because the glass platters shattered and now its just a happy shaker box with a tail. If you want failure rate and slowness, the record players can't be beat. SSD's are still new but evolving very very nicely and myself only look forward to the record playing HDD's to be on a techs wall for historical purposes by 2024. Also if she wants to learn the truth about drives she should work at a computer repair shop for a few years for experience instead reading peer reviews and marketing material.
 
carnage0651 said:
You left out that tape drives from back in the 1990s still have more capacity than most HDD's today. If a company wanted to, it could have kept innovating that type of data storage, and we'd all have tapes in our computers. The world just went a different way..
Click to expand...

Really ? I am intrigued... what tape system from the 1990s has more storage than a single 3.5" hard drive nowadays ? Keep in mind, We now have 4TB 3.5" drives, and the largest storage tapes back in the 1990s were the DDS3 tapes that stored 4GB each, and only got to 8GB if you compressed the entire volume. The only tape storage systems I know from back then that would be capable of 4TB of storage, would be an enterprise level tape robot that too up an entire 7' tall rack, and cost around $300,000. So what is this magical tape that you know of that could store 4TB on it back in the 1990s ? Checking the market nowadays, the largest retail accessible tapes are just reaching 4TB as well, and they cost thousands of dollars for a single drive. Please enlighten me, because I would be very interested in a tape drive capable of hundreds of TB, or even dozens of TB.
 
Uhh graduate student says...They neglected to talk to undergrads and high school kids then they could have written : Multiple researchers agree...

I remember reading about some real hmm grownup scientist who at the end of 19th century was claiming that the earth would be covered by 9 feet lawyer of manure mainly from horses by 1950. The logic was that population would increase and proportionally more horses would be needed. Of course he failed to account for internal combustion, public transportation, 2 world wars and many other things.
 
I fail to see why Hard Drives don't have larger read/write heads; or better yet, rather than have a single mobile actuator move to-and-fro, just make a fixed single (or multiple) long strip or radial sensor that goes the length of the inner to outer drive. This way, you remove the whole motion entirely and you can read/write along the whole strip. The drive can still spin at whatever RPM you want and now you've increased the IOPS by a magnitude of times. Why have they not done this already? This is something I thought about the first moment I cracked my first ESDI drive open in the early 90's. It seems stupid that all this data is read from a single sensor on an actuator.
 


Thats a good thought... how many tracks could the platter have? 1 head per track mounted on a fixed arm. No need for a moving actuator and the ability to read and write to all tracks at the same time. Seems so simple.

I would assume some control logic would be needed to compensate for the inside of the platter spinning slower than the out side, and for the ability to control multiple read heads at once. But it does not seem to be that much more complex than whatever control systems are in place on current drives.

Maybe the cost of multiple read heads?
 
If you are all so smart why dont you reply to her here? lgrupp@cs.ucsd.edu
 
There is a tape drive that was mostly an experiment to see how high they can go... it was the size of a car and length of nearly a mile... That had the capacity of 15.6TB this was of course never used because nobody in their right mind thought ANYONE would need this much storage. they didn't even use TB they used MB to measure it. Not sure of the date but I'd say it was late 1980s
 
Having been in the PC business for over 20 years, anyone who tells me where tech will max out in 5+ years is foolish. Going out to 2024 is more so. How about optical and magnetic drives go obsolete when holigraphic technology is discovered in 2018? HA!
 
Do you remember what devices and technologies were used 12years ago?FDD,MD,CD-RW...craps!
12 years ahead!?NAND flash SSD will be crap,too.
 
2024... Honestly I would be surprised if most of what I'll be cramming in my new computer shortly will not see major hardware change; or complete replacement by a superior tech over a 12 year period. ie; for simpletons; Jump between super nintendo and xbox over a 12 year-ish period. Here's a question will I still be putting gas in my car in 2024.
 
Freggo: In 1966, a removable hard disk weighed about 2 lbs for each Meg in IBM's 5 and 10 Meg drives, and had to be hefted all day long into and out of IBM Hard Drive Enclosures as big as modern dishwashers. What we now have was my idea of the far distant future's unbelievable computer wonders.
 
If size is an issue, why not just make SSDs the same size as normal hard disks?
 
[citation][nom]stingstang[/nom]You left out that tape drives from back in the 1990s still have more capacity than most HDD's today. If a company wanted to, it could have kept innovating that type of data storage, and we'd all have tapes in our computers. The world just went a different way..[/citation]

Try accessing data in the middle of the tape...
Tapes are ok for backup of course, but as a live online storage medium they are basically useless.

 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom