SSD's are Doomed: According to Researchers.

Doomedspeed

Doomedspeed

Doomedspeed

Soldato
Joined
13 Oct 2011
Posts
11,881
Location
Melbourne, Australia
Researchers at the University of California, San Diego, have concluded that solid state drives (SSDs) have a bleak future in the evolution of computing technology. They have discovered that fast flash based storage are facing come pretty glaring technology hurdles during their natural course of evolution, which they don't think it will overcome. To begin with, shrinking (miniaturizing) them, to increase capacity or decrease manufacturing costs, will severely degrade performance beyond a point, 6.5 nm silicon fab process.

The scientists studied 45 different flash chips in various sizes, which showed that scaling of latencies and error-rates are 'tolerable' enough as the technology miniaturizes only till 6.5 nm, or the year 2024, when this fab process will be common, beyond which they question the drives' viability. Beyond this point, the more capacity you squeeze into flash memory chips, the more performance degrade (latency and error-rate scale beyond tolerable scales).


147a.jpg

147b.jpg


While the density of SSDs grows and the cost per gigabyte shrinks, "everything else about them is poised to get worse," said Laura Grupp, a graduate student at the University of California, San Diego. "This makes the future of SSDs cloudy: While the growing capacity of SSDs and high IOP rates will make them attractive for many applications, the reduction in performance that is necessary to increase capacity while keeping costs in check may make it difficult for SSDs to scale as a viable technology for some applications," the author of the study.

The study, entitled "The Bleak Future of NAND Flash Memory", can be accessed here.
Click to expand...

I don't agree!
 
Damn laws of physics! always getting in the way of our fun.

I don't think anyone expected SSD performance/affordability to keep increasing forever - since like CPUs it is limited by manufacturing processes of transistors, which when getting this small start running into new and interesting issues.

It is conceivable if these issues are insurmountable we will be in a similar position to mechanical hard drives - where they plateau. However this did spur on the development and adapoption of an entirely different technology which is now becoming mainstream (SSDs). Who's to say that something similar won't happen when SSDs start to plateau.

That said, who knows - maybe in the next 12 years they will get round this roadblock - perhaps using a different material for semiconductors, so instead the roadblock is put back a few years (and a few generations of SSDs).
 
They are in no way taking into account the fact that we will most likely have a new method of data storage that isnt NAND memory by 2024.
 
Maybe I'm getting old, but... is this really a problem?

I mean, the big ticket items on most hard disks would be audio, video, and maybe games. I wouldn't imagine there being any real performance benefit to having a 4gb video on SSD over a 3TB mechnical.

If SSDs can get up to say, 1TB, that'd store Windows, apps, and any documents... as far as games go, there might be some arguement for making them so that the game files install on a SSD but all the cut scenes/video/media go onto HDD.
 
sehcure said:
They are in no way taking into account the fact that we will most likely have a new method of data storage that isnt NAND memory by 2024.
Click to expand...

Of course they aren't... you can hardly measure and extrapolate the performance of something that hasn't been invented yet :p

It's merely a look at how the current implementation of SSD are going to reach a 'max theoretical' limit at 6.5nm.

And that by that point someone will need to have come up with an alternative.

Which they no doubt will have.
 
They said the same about silicon sizes as well 2-3 years ago, something about going smaller than 16nm introducing quantum tunneling effects and standard lithography not being able to cope. In the years since, Intel say they have a feasible road to 11nm so they must've solved those particular problems...

Although of course 6.5nm is pretty small, so maybe it will be an issue.
 
Considering a 2.5" drive is around a sixth the volume of a 3.5" drive you could have a 6 TB+ SSD in a 3.5" desktop format if you could justify the price. :)
(Probably larger as the volume of the 6 2.5" included 6 bodies)

With 20nm NAND flash (2013) that could be doubled to 12 TB+. :)

This probably won't happen due to cost, but in theory it's possible.
 
They do say in the study that this is a theoretical study.

It doesn't involve the black magic that modern SSD controllers currently do, and will probably do more effectively by 2024.
 
Sasahara said:
Considering a 2.5" drive is around a sixth the volume of a 3.5" drive you could have a 6 TB+ SSD in a 3.5" desktop format if you could justify the price. :)
(Probably larger as the volume of the 6 2.5" included 6 bodies)

With 20nm NAND flash (2013) that could be doubled to 12 TB+. :)

This probably won't happen due to cost, but in theory it's possible.
Click to expand...

The irony being of course, that SSD has the potential to overtake Mechanical drives in terms of capacity over the next couple of years - if the volume increased significantly we might even see prices coming down enough to make them more affordable - perhaps it's the mechanical platter that's going the way of Tapes ;)
 
Storage densities will have increased to the 10p/gb point before we reach 6.5nm, making it affordable to keep even your steam directory on SSD. Mass media is pretty much a non issue, we'll all be streaming on demand or keeping it in the cloud by then. Average requirements for home storage needs are going to go down over the next decade, not up.

That said, I still expect alternative solid state solutions (MRAM etc) to properly emerge before 2024, which should kill off flash.
 
error-rates are 'tolerable' enough as the technology miniaturizes only till 6.5 nm, or the year 2024, when this fab process will be common, beyond which they question the drives' viability
Click to expand...

YEa, thats a major downer- technology will obsolete in 12 years time... Who cares about such projections, 12 years is like a century in this business.
 
Borsch said:
YEa, thats a major downer- technology will obsolete in 12 years time... Who cares about such projections, 12 years is like a century in this business.
Click to expand...

While I agree that the tech will most probably evolve beyond this issue, it is worth bearing in mind that physical limits, and more importantly design limits can hamper future growth. Some of us are old enough to remember the 640kb issue, not to mention the issue with 2 digit year groups! What is reassuring is that people are aware of the limits now, not 6 months before they are reached!
 
The gates 640k quote comes to mind here.

"I don't see how we can possibly overcome this super advanced SSD technology hurdle in the future. The technology just isn't there" :rolleyes:
 
Pretty boring, nand is already done for in so much as, newer types of memory all look seriously better already and coming up soonish. IE higher storage capacity and the ability to write not whole blocks at a time.

Either way, who cares, an ssd now will gives performance that a HDD will NEVER achieve, sequential is at this point in time, almost worthless, the thing that "makes" an ssd what it is and makes random small file performance and high queue depth performance so good is....... essentially no latency. HDD's will never overcome this. Likewise there isn't actually any reason you need 2tb ssd's, because the vast majority of people's storage needs aren't performance related.

You have 30gb OS which eventually should shrink rather than grow(you can already trim down a vista/7 install to a fraction of that but cause long term problems, if only MS would make a tiny version of it that actually worked), a few games and all your video, music, work on an HDD, no need for "performance" to playback a simple high def file at a fraction of the speed even a HDD is capable.

So the situation is, if in 10 years you can get a 256gb ssd for £40, and a 4tb ssd for £400, and a 4tb hdd for £40.... the vast majority of people will get a 256gb ssd and a 4tb hdd anyway. There is simply no requirement now for all storage to be ultra fast, nor will that change in the future.

Anyway, the vast majority of performance from chips is from having more chips and accessing them all at the same time. Marginally increased latency won't make a significant difference if you can access 4 times as many chips, which will end up with vastly more performance anyway.
 
a quick copy and paste
The memory capacity of electronics devices could be increased in future thanks to an organic data storage system using ternary rather than binary data storage. The current prototype is designed for permanent data storage, and can be written once but read multiple times, but the Chinese researchers hope to develop re-writable data storage based on the technology.
 
You must log in or register to reply here.
Back
Top Bottom