Slashdot Mirror
Why SSDs Won't Replace Hard Drives
storagedude writes "Flash drive capacities have been expanding dramatically in recent years, but this article says that's about to change, in part because of the limits of current lithography technology. Meanwhile, disk drive densities will continue to grow, which the author says will mean many years before solid state drives replace hard drives — if they ever do. From the article: 'The bottom line is that there are limits to how small things can get with current technology. Flash densities are going to have data density growth problems, just as other storage technologies have had over the last 30 years. This should surprise no one. And the lithography problem for flash doesn't end there. Jeff Layton, Enterprise Technologist for HPC at Dell, notes that as lithography gets smaller, NAND has more and more troubles — the voltages don't decrease, so the probability of causing an accidental data corruption of a neighboring NAND goes up. "So at some point, you just can't reduce the size and hope to not have data corruption," notes Layton.'"
Yeah, there's NO way that SSD technology will somehow evolve further than it has till now. It's after all SEVERAL years old by now!
With SSDs, I'm sure there is always another axis of improvement, similar to with CPUs, when you hit a wall with them, go SMP. When SMP doesn't scale, crank up the clock speed, etc.
What I wonder is what can be focused on to make SSDs be able to store more. We can always stick more chips in an enclosure, and the cooling needs for SSDs are far less than the cooling needed for CPUs.
Was plenty for my needs and boots Ubuntu in 20 seconds. Barely uses power when not in use. I'm a believer.
While the reasoning is interesting, and valid for all I know, why are we trying to say some bit of technology isn't going to work out ever? What's the point? Either it won't work out and that will be something the market will handle independent of whether you foresaw it or not, or a solution will be found and you'll just be wrong.
I'm reminded of an Arthur C. Clarke quote: "When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong."
If will be a long time before development of the horseless carriage will overtake the technology of my steam-powered ornithopter!
Give a man a fish and you have fed him for today. Teach a man to fish, and he'll say "WHERE'S MY FISH, YOU IDIOT?"
The bottom line is that there are limits to how small things can get with current technology.
They're right, SSDs won't replace hard drives with the current technology. If only we had a way to improve technology over time!
How far does the storage capacity really need to expand? Hard drives are in the terabyte range now, but not many people really use that much. On media servers or something, maybe, but on your average computer? I've got 50GB in my laptop once you account for my windows partition, and I'm fine with that. A 320GB SSD would last me a lifetime, especially considering the btrfs is supposed to support on-the-fly compression.
Like I said, the only place where I can see the large capacities being needed is behind the scenes on a server or similar device, in which case hard disks aren't much of a problem. On consumer computers, I'm pretty sure they're going to catch on.
Funny may not give karma, but +5 Informative never made anyone snort coffee out their nose.
"Science reporting organizations have been expanding dramatically in recent years, but this article says that's about to change, in part because of the limits of current literacy education. Meanwhile, tabloid reporting will continue to grow, which the author says will mean many years before solid science reporting replaces sensationalism -- if they ever do. From the article: 'The bottom line is that there are limits to how smart things can get with current society. Universities are going to have student density growth problems, just as other societies have had over the last 30 years. This should surprise no one. And the literacy problem for journalism doesn't end there. Buff Clayton, Editor in chief for The Onion at Delaware, notes that as literacy gets smaller, science reporting has more and more troubles -- the bullshit PR releases don't decrease, so the probability of causing accidental sensationalism goes up. "So at some point, you just can't reduce the literacy and hope to not have reader confusion," notes Clayton.'"
It takes a man to suffer ignorance and smile
Be yourself no matter what they say
SSDs already are large enough... for normal people. 1TB is here, and it's WAY overboard - most people can't even use 256GB. For the average user's needs, 64GB is perfect for today's OSes. The article's claim is laughable.
It's economic feasability, too. Rotating media is roughly $100/terabyte, it's gonna take more than one breakthrough for SSD to come close to that.
Nifty new technology doesn't get bought because it's nifty-new, it gets bought because it fills the need better than its predecessor for the price.
And YES there are plenty of applications where multiple terabytes are necessary, maybe not on your home system.
In case you're wondering, I have both on my system: / is SSD, /home is multi-terabyte RAID. Rotating mechanical media is sticking around at least for now.
How many of these could you fit in the space of a standard HD case?
I know, someone's gonna lecture me on how this isn't at all a fair comparison...
Correct me if I'm wrong here - and I usually am wrong - but aren't we "limited" now only by controllers and the *price* of the NAND chips? I've read anandtech's last few SSD manifestos and it seems the controllers' speeds and the price of the NAND - not really anything else - is limiting their absolute capacity. I recall engadget doing several reviews of SATA and PCI-E SSDs with capacities up to 1TB. Granted the 1TB Z-drive was between $1,500 to $2,000 back in March of 2009, but you get the idea. We can make a very large SSD today. It's just not affordable.
To wit, who honestly has a larger than 1TB disk inside their machine right now? I'd imagine not terribly many, as a percentage of all computer owners. Indeed at home I have twin 700-ish GB Caviar Blacks in a RAID 1 configuration, of which I'm using maybe 30% of their capacity.
TFA doesn't actually make any arguments about price directly. It indirectly suggests price of the drives is related to lithography resolution, but provides nothing to back that up.
It seems to me that over time as yields on current technology increase and fab costs are recouped, the price of current technology will go down.
So if we can make a 1TB disk today, it'll be the same 1TB disk in a year or two, except less expensive, probably faster, and probably more reliable.
Some context would be nice. It may be that SSDs end up replacing conventional hard drives on, say, all laptops. Or all personal desktops that don't also double as servers. Or we may see a two-tier situation develop where SSDs are used for day-to-day operations in the enterprise and hard drives used for storing backups, or storing infrequently accessed archival data.
As HDDs continue to fail before their expected lifetimes due to sensitivity to movement and the general worse state of moving parts vs unmoving parts, people may start to flock towards SSDs as replacements, especially as people start to notice the many benefits of SSDs over HDDs. They'd have to realize though that extraordinary wear could shorten the length of an average MLC and that SSDs even on normal usage are not meant to last forever, but with the improvement to wear leveling this may be less of a problem in the future.
Yes, we always do. Don't underestimate the space needed to store pr0n.
SD cards go up to 32GB. They're only 2.1mm thick. Just piling them up you could fit 200 or so in the volume of a typical hard disk, and it's not like an SD card is the most space efficient means of storage since a lot of the volume is taken up with the plastic case. Micro-SD can go to a quarter of the capacity in a tenth of the size. So we can squeeze at least 16TB into the same volume. That's probably adequate for a typical home user. The price is the issue here.
Or, more to the point, do we need to store all of our biggest files (media, usually) on a SSD? I, for one, have no problems playing music, looking at photos, and watching movies on my normal hard drive. I have a SSD and a traditional HDD in my computer. I use my (much larger) HDD for storing my media, and my SSD for storing high-traffic things like my OS and games. I get the speed I need for my applications, and the size I need for my media.
http://www.maximumpc.com/article/news/breakthrough_nand_flash_memory_could_lead_10gbs_ssd_writes We have had a breakthrough in solving the voltage problem. I think the authoer is nothing but idiotic to believe that SSD isn't going to replace hdds for the average consumer. Later this year intel is going to release its G3 SSDs, with the lArgest at 600GB. G2 drives were 60% cheaper than G1 drives. Let's hope we see a similar drop.
The problem with "online" storage is that you can end up offline.
When that happens, the fact that my phone can hold my entire music collection is a handy thing.
It always amazes me when people talk about the cloud as if all of the necessary network infastructure was already there. It's not. Mobile networking is CRAP and mobile networking providers seem intent on also making it EXPENSIVE too.
It's the cloud that sucks. SSDs have potential. Their main problem is that they're terribly expensive. They are not likely to overtake spinny disks any time soon because of this.
A Pirate and a Puritan look the same on a balance sheet.
Correct me if I'm wrong here - and I usually am wrong -
I'm usually right... but that statement might be one of the exceptions. :)
TFA doesn't actually make any arguments about price directly. It indirectly suggests price of the drives is related to lithography resolution, but provides nothing to back that up.
It seems to me that over time as yields on current technology increase and fab costs are recouped, the price of current technology will go down.
It's a basic maxim of the silicon industry that cost is directly proportional to die area. To simplify, you can consider the silicon fab to have a fixed cost per wafer. Therefore the more die fit on a wafer, the cheaper each chip becomes. The two main ways to do this are by reducing the amount of functionality on each chip (undesirable when the goal is to increase capacity), or to move to a smaller lithography so you can fit many more die on a wafer. While new lithography generations have frequently allowed greater performance, even if they don't they are deployed anyway because it reduces cost for the manufacturer.
Yield improvements and paying off R&D both will help cost, but only to a limited extent. Yields for a production lithography should already be quite high and will asymptotically approach 1. Once R&D is payed off the cost will drop, but there still remains a very large fixed cost per wafer. Neither is going to come close to the cost benefit of being able to, say, go from a 45nm to 32nm process and get roughly 40% more die per wafer.
So yeah price will come down for other reasons, but in the long term price reductions in flash memory devices are going to depend on using smaller lithographies just like it does for other semiconductor devices. The author probably just didn't think to explain this aspect of it, since it's such a well-known aspect of the silicon industry.
On the other hand, people were saying that CMOS processes used in CPUs were going to reach fundamental limits 20 years ago. And 15. And 10. And 5. And oh sure, some of those limits were reached, but then clever people worked around them. The statement in the article amounts to "We can't just blindly reduce lithography size without changing anything else indefinitely", which is true but also kinda pointless since the people working on smaller lithographies for flash are probably aware. In the end exponential progressions like this can't last for ever, but I'm not about to tell the process engineers that they aren't going to be able to find enough tricks to keep it going long enough.
The enemies of Democracy are
This is what annoys me is that it seems like Flash is idea as a cache for magnetic HDDs. The same principle is already at work in our CPUs:
So a modern CPU is way faster than modern RAM. The access times are much lower. How then, can we have a system not hamstrung by RAM? The answer is cache. With a good system of high speed L1/L2 (and sometimes L3 cache) we can have our cake and eat it too. You have a few megabytes of expensive high clock SRAM right on the core. You have a few gigabytes of cheap DRAM clocked much slower. With proper caching, you then get 90-95% of the expected speed of the SRAM. Nearly all of the speed, a fraction of the cost.
Why not HDDs then? Have the RAM on there (L1) and a couple gigabytes of flash (L2) pared with the disk. Use an intelligent caching algorithm (as in not just the first part of the drive) to cache reads and writes. This should again offer most of the expected speed of the flash, while still offering a low price.
I'd pay for that. Say a full magnetic drive is $100 for 1TB. A full SSD is $3000 for 1TB. A Hybrid 1TB drive, which features 4GB of flash, is $200 but performs 50% faster than the magnetic drive and deals with simultaneous reads and writes much better. I'd buy that.
Unfortunately all the hybrids are for laptops and use it to save power, not to speed things up.
You mean like this
Chips can be super tiny, spinning platters aren't so much. It wasn't that long ago that people were predicting hard drives couldn't get much more dense due to neighboring bits flipping each other too frequently, then they had a breakthrough with perpendicular recording. Look at what a 4 gig microSD card looks like. Then imagine how many of those could fit in a 3.5" drive chassis if carefully stacked them in there. I bet you'd be able to stack more than 1000 easily. Aside from that, even if each format may not be able to scale down significantly further, I can only collect so many TB of "linux isos" before I can't find a damn thing.
The hot new solid state non-volatile memory technologies are phase-change memory (PRAM), memristors, ferroelectric RAM, resistive RAM.
Some of these technologies are much more area-efficient than Flash, and will stack in pseudo-3D chips reasonably well (memristors in particular should stack in full 3-D arrays very efficiently...).
The general observation that disks have the lead right now is true, but the other technologies close a lot of the gap, and the growth curves look very similar after that. Who knows if it ever gets cheap enough to completely replace disks in our lifetimes, but there is hope of seeing that.
That does entirely change the game on system architecture. Disks are slow and far away from the CPU. Solid state memory can be as close or nearly as close as DRAM, and if it doesn't require a lot of handholding on lifecycle management (wear rates etc - Flash is horrible here) then can be used and managed as a simple byte or block array rather than the whole "filesystem" crap we now use. We still may want POSIX like abstractions for parts of storage management, but life is so much easier if the back end store is just a block array we read/write than if it's really a spinning disk, behind a cache, behind a controller, behind a SATA/SAS bus, behind a controller, behind a PCI bus, behind a southbridge, ....
I agree. I thought this was going to be an article about something new instead of just throwing in the word lithography into a tired old argument.
I bought an SSD after by HDD overheated and died in my netbook. I bought it on ebay because I was trying to be stingy. Turns out it wasn't a good choice. It runs pretty darn hot and doesn't reduce power consumption that much. It is not appreciably faster. The worst part is it isn't universally well known which is important with linux. I often get Grub Boot error or other system freezes for no apparent reason. Also it can be really slow for some write heavy applications. Overall I wouldn't do it again. The only good aspect is that at only 32GB by collection of downloaded nonsense can't get too big or out of control so I don't have to worry about backups as much because it isn't so big... and I get more work done with less movies on there to watch.
What I really want is a cooling mechanism for a normal HDD that sits inside the case (not a bulky laptop pad). I really don't think it is acceptable that the disk burns up under normal usage on a hot day. Things are just packed too tight and I have yet to find any after market add-on I can use to keep the temperature reasonable. I always have fans in my desktop because I know temperature is inversely proportional with HDD lifespan and I don't enjoy disk failures.
SSD was a real dissappointment for me. I was willing to trade storage capacity for the supposed benefits of cooler, less power, quicker, less likely to fail. But the disk I bought doesn't seem to deliver on any of those fronts in any signifigant way and the reliability is indeed less: with a spinning disk I know the typical problems and solutions and ways to make the most of it. With the SSD I have no idea why the system periodically freezes, or if it is trustworthy given the whole weirdness about having SSD specific filesystem requirements that I can't decide which method would be a wise so I stick with EXT4 and hope things work out.
Stupidity is its own reward.
if you're building a new one, it shouldn't be much of a bother to fit a physically bigger drive inside your case.
Unless I'm building a laptop, or an all-in-one, or a slim PC to put next to the TV, etc. Not every PC is a traditional tower.
If only we had a way to improve technology over time!
Compare not next year's SSDs to today's HDDs; instead compare next year's SSDs to next year's HDDs. If both SSDs and HDDs improve at the same rate over time, HDDs will keep their lead compared to SSDs for any application that isn't handheld, vibration-sensitive, or seek-heavy.
The author finds "some good data on Wikipedia" (respect!) showing that the "lithography size" will be reduced from 32nm in 2010 to 11nm in 2022. He calculates this to be a "volumetric improvement" of 50%. There I was thinking that it was an 846% improvement, but I hadn't taken the third dimension into account.
Nevertheless, I think the author has a point, but he is missing part of the picture: NAND flash SSDs may not replace HDDs any time soon, but other types of non-volatile memory may well do so.
HDD densities will probably increase, but the slow access and transfer times and the static unrecoverable error rate will probably relegate them to use for back-ups or as cheap mass-storage devices for non-critical data. SSDs, however, are not restricted by the limits of NAND flash. Non-volatile memory technologies such spin transfer torque RAM and phase-change RAM have a good chance of replacing NAND flash memory in SSDs. These technologies are available today. Memristors are probably the most exciting development, as they promise a breakthrough in memory density. HP have a memristor-based design that could make petabyte SSDs possible, but we'll probably have to wait a few more years to see if that pans out. There are also major advances being made in fabrication technology, with cheap "printable" electronics already in consumer devices.
Real random-access memory that is cheap, reliable and fast is probably only a few years away from the mass market. There is so much money to be made by such an advance that R&D spending will not be lacking. So, the author is wrong; SSDs will dominate in the near future, just not NAND flash SSDs.
P.S. I don't have any SSDs because they are too small and expensive compared to my 1TB HDDs!