Samsung 256GB SSD is World's Fastest

i4u submitted one of many holiday weekend slow news day stories which starts "Samsung Electronics announced today the world's fastest, 2.5", 256GB multi-level cell (MLC) based solid state drive (SSD) using a SATA II interface. Performance data of the new Samsung 256GB SSD features a sequential read speed of 200 megabytes per second (MB/s) and sequential write speed of 160MB/s. The Samsung MLC-based 2.5-inch 256GB SSD is about 2.4 times faster than a typical HDD. Furthermore, the new 256 GB SSD is only 9.5 millimeters (mm) thick, and measures 100.3x69.85 mm. Samsung is expected to begin mass producing the 2.5-inch, 256GB SSD by year end, with customer samples available in September. A 256GB capacity is getting large enough to replace hard-drives for good — now just the prices just need to come down further for large capacity SSDs."

Comments

Random write ops?

[Fweeky]

So, can this one push more than 20 random writes per second?

Re:Random write ops?

[Anonymous Coward]

yeah... right.

Re:Random write ops?

[Kjella]

Uh, the last SSD review I saw that had considerably worse specs than this just barely came behind the VelociRaptor in the random write tests. Unless you've got a special disk trashing benchmark to make SSDs look bad, I'm fairly sure this wins on all counts. In fact, 20 random writes/sec sounds more like trolling than insigthful to me...

Re:Random write ops?

[Anonymous Coward]

What? Random access has never been an issue with SSDs - it's where they shine and HDDs look like shit. Sequential access is where HDDs have remained king until now.

Re:Random write ops?

[Fweeky]

Every benchmark I've seen on SSD's have shown random IOPS of between 20 and 120/sec, ranging between cheaper consumer drives and more expensive enterprisey models; writing single blocks to random locations completely demolish their performance because such small writes often require the drive to erase huge blocks.

New techniques try to avoid this by basically turning random writes into sequential ones; once you've erased a 4+MB block, you put all new writes into that block (you can turn a 0 into a 1 without an expensive erase cycle) and remap it similarly to how it's done with wear leveling. I'm not aware of anyone actually doing this yet, though.

Re:Random write ops?

[Fweeky]

No, SSD's have always shined at random *reads*. Small random writes have traditionally been where they're very weak; you might manage 160MB/s writing large chunks, but if you're droping 16k blocks all over the place (as, e.g, databases are apt to do) you'll be lucky to manage 1MB/s because of the overhead each write incurrs, certainly on cheaper drives aimed at portable use.

Hence, it's a perfectly reasonable question; depending on how they've implemented it, they could be anywhere from 20-20,000 random writes/sec.

Re:Random write ops?

[GreyWolf3000]

No, SSD's have always shined at random *reads*. Small random writes have traditionally been where they're very weak; you might manage 160MB/s writing large chunks, but if you're droping 16k blocks all over the place (as, e.g, databases are apt to do) you'll be lucky to manage 1MB/s because of the overhead each write incurrs, certainly on cheaper drives aimed at portable use. Hence, it's a perfectly reasonable question; depending on how they've implemented it, they could be anywhere from 20-20,000 random writes/sec. To expound, NAND flash chips are broken up into 128KB 'blocks' which in turn comprise 64 2KB 'pages.' You can read any page you want on the entire chip in the same amount of time (no moving parts), but to *write* any particular page, you need to perform an erase on the *entire* block. Let me repeat, to write one page to NAND flash (2 KB), you have to erase a 128 KB block. The reset operation will transition all bits on that block to a 1, and you have to go back and tell it which bits to set to zero.

Re:Random write ops?

[TheRaven64]

Unless flash has changed a lot since I last looked at it, this isn't quite true. To write a 2KB page, you first need to erase a 128KB block, but you can then write into the remaining 126KB without needing an erase first. If you're using something like LFS or ZFS on your disk, this translates to very fast writes. If you're using a filesystem that doesn't have copy-on-write semantics, then this will still be quite slow because you will rarely encounter this kind of access pattern (you will be more likely to read 128KB, modify 2KB in a buffer and then re-write 128KB, which, as you explained, is very slow since it requires a 128KB read followed by a 128KB write just to write 2KB - the same reason that RAID-5 is slow for writes).

Re:Random write ops?

[GreyWolf3000]

What you said doesn't contradict what I said at all. You can definitely achieve much better performance by not scattering your writes; sequential access is very fast. For random writes, you're going to have to seriously buffer to avoid wasting erase cycles. Fortunately, the same algorithm that avoids erase cycles for performance is also the same algorithm that does wear leveling. I don't know too much about ZFS or how copy-on-write would help except on a purely abstract level.

Re:Random write ops?

[Orange Crush]

Isn't that the other way around? I thought flash memory's "blank" state was all 1s?

Re:Random write ops?

[Anonymous Coward]

[O]nce you've erased a 4+MB block, you put all new writes into that block [...] I'm not aware of anyone actually doing this yet, though. Some filesystems (for example ZFS) will do this.

Re:Random write ops?

[TwilightSentry]

NOR flash clears to 1 and can have individual bits set to 0. NAND flash clears to 0 and can have bits set to 1.

Re:Random write ops?

[jgoemat]

Set your filesystem to use 128kb blocks, problem solved :) The ZFS or copy-on-write helps because your old blocks aren't written over, all changes are written to new areas on the disk. Let's use A-Z for 128k blocks and 00-3F for the 2kb pages in them, so you have 3,328mb. Let's say that A00-M3F are used. Now let's say you do random writes to data at A13, B27, E1F and L19. Normally you would erase blocks A, B, E, and L and re-write 4*128kb or 512kb of data. With copy-on-write you simply erase (if not already erased) block N and write to N00, N01, N02, and N03. That's one erase and 8kb written instead of 4 erases and 512kb.

Re:Random write ops?

[Impy the Impiuos Imp]

> the new Samsung 256GB SSD features a sequential read speed of 200 megabytes per second (MB/s)

The article continues:

"That is at the hardware level. Unfortunately, some OS designs prevent reads any faster than 640k/s, 'since 640k per second is all the read speed anybody will ever need.' "

Summary

[Mark Trade]

Don't buy any other similar products. Ours will come out Really Soon (TM). At least we hope so.

Re:Summary

[TheRaven64]

And since it's using MLCs, expect to buy another one quite soon after. While SLC flash is up to the 100,000 to 1,000,000 rewrites level, MLC is still closer to 10,000. This, combined with the larger cell sizes on most MLC products means that it is likely to wear out much faster. This is why most flash manufacturers are only pushing MLC technology for consumer electronics type applications (e.g. MP3 players), where you want high density and low cost but don't perform many writes.

Seems like the complexity is lower

[Eccles]

Looking at a hard drive, it's got lots of moving parts, the need for sealing, etc. One would think that in the long run a solid state drive that is just a few chips and connecting logic would be cheaper to produce once you have the facilities.

Re:Seems like the complexity is lower

[__aamnbm3774]

you're probably right.

but just like CD's are cheaper to produce than cassettes, that doesn't mean the cost will ever come down.

Re:Seems like the complexity is lower

[Kjella]

Easier to miniaturize, certainly. Right now they're doing massive RAID0 to get performance, I wonder what it'd be like if they could do RAID1/5/6 for example - forget hard disk crashes more or less, just replace some flash plug-in modules in your SSD. Ok the electronics could still fry, it could get lost or stolen but mechanical failure seems to be the typical killer.

Re:Seems like the complexity is lower

[kipman725]

they have come down a blank cd can be had for less than a pence if you buy in large enough quantities. It's just not viable to sell CD's at less than the cost of transportation in individual quantities.

Re:Seems like the complexity is lower

[__aamnbm3774]

Then explain how cassettes were selling less than CD's when they were first introduced?

Over time and with current gas prices, I can comprehend 17 dollar CD's, but I'm talking about before the MP3 era and before any type of internet radio. They took advantage of the excited market to keep music prices inflated and rip us off. (when we didn't have a choice)

Please quit your job at the RIAA.

Re:Seems like the complexity is lower

[Anonymous Coward]

Sometime in the future SSDs may become chaeper to manufacture than HDDs but for the near future they will remain more expensive.

In a hard drive you only need one of the expensive, complex parts (i.e. read read) and the rest (platter) is very cheap. As the capacity is defined by the amount of platter, increasing the capacity doesn't increase the cost considerably.
This is in contrast to SSDs where each bit is stored with integrated circuitry of moderate cost. To increase the capacity you need more transistors, and hence the medium doesn't scale as well as magnetic HDDs.

Re:Seems like the complexity is lower

[vux984]

but just like CD's are cheaper to produce than cassettes, that doesn't mean the cost will ever come down.

When were you ever able to buy 100 blank cassettes for 20$?

Re:Seems like the complexity is lower

[canuck57]

Looking at a hard drive, it's got lots of moving parts, the need for sealing, etc. One would think that in the long run a solid state drive that is just a few chips and connecting logic would be cheaper to produce once you have the facilities.

Given sufficient amount of time, solid state SSD will likely overtake hard drives. But I think many industry analysts are far too quick to estimate wide adoption if the SSD media over hard drives. It will be slow. And I have heard those predictions 10 years ago.

Problems exist in SSD adoption, 3 huge ones.

• density, hard drives seem to be always many steps ahead in density
• costs, a $100 hard drive w. 1TB version the SSD 1TB cost?
• write speed/reliability issues with SSD

Oh, the SSD will creep in, but I don't expect it to wipe out hard drives any time soon. I will say when the 640GB SSD is under $250 it's adoption will soar for laptops. By that time, the 1TB mechanical hard drives will be under $100 and 2 or perhaps 3TB drives may exist as well. But we are optimistically 4-5 years from this point. For the data center, even longer. The write and cost issues must be totally resolved for that, as some drives in busy systems go nuts on writes and can't afford a hit. If I want to buy 10PB of storage, and SSD is twice the price, it will loose.

What you might see in widespread adoption first is say affordable 64GB versions of SSD for the OS, and a adjunct 1TB hard drive for raw storage.

Re:Seems like the complexity is lower

[PIBM]

You can have 100 cassettes with non-free content for less than 20$ at any (old) yard sale!

Re:Seems like the complexity is lower

[rainhill]

Ideally, one sould be able to plug in these SSDs, just same as a RAM module.. perhaps a redesign of motherboards are needed.

Re:Seems like the complexity is lower

[donaldm]

Compared to a hard drive the Solid State Disk has no moving part however they don't have the reliability of modern quality hard drives but if you look back to the 1970's and 1980's disk drives were fairly small in capacity but large and relatively power hungry they were also expensive and not that reliable. The problem with hard drives even today is that they have moving parts which will eventually fail. Hence the Mean Time Between Failure rating which are sort of a rubbery figure that gives a guideline on how reliable the device is.

The MTBF of this solid state drive is supposed to be 1,000,000 hours which is actually not a bad figure since that translates to approx 114 years. Even taking say 5% (my rubbery figure) that still translates to approx 5 years which if you look at many laptop disks that is reasonably reliable.

I doubt that SSD's will replace cheap high capacity hard disks any time soon but their main use at the moment would be in laptops and any device that is requires low power consumption and possibly a greater resistance to vibration.

Like anything over time prices will come down with reliability and capacity going up. To determine how far a product will drop in price you actually need to know the overall production costs of the product which not only includes fabrication but packaging, delivery and marketing. An interesting example is flash disks which have dropped significantly in price but increased in capacity with the lower capacity line actually dropped instead of continuing to drop the price per lower capacity drive.

42 zillion dollars?

[billcopc]

I like the idea of the SSD, but I think they need to concentrate on lowering costs down to earth before flaunting their capacity achievements. Hell, any monkey can build a 500-TB mega-RAID stripe with a large enough budget.

When this SSD is cheap enough that I can buy 3-4 of them and stripe that into a bus-raping powerhouse, for less than a mortgage payment, then we'll talk.

Re:42 zillion dollars?

[Sascha J.]

If the SSD's price only increases very slightly with greatly increasing capacity, they already lowered the cost.

Also, it doesn't help to have cheap 32GB SSDs when nobody buys them and you can't really launch into mass production because you are stuck with a niche market. To drive down the price you need to be able to produce them en mass and in order to do that you need to catch up (or outstrip) existing technology.

Re:42 zillion dollars?

[Anonymous Coward]

When this SSD is cheap enough that I can buy 3-4 of them and stripe that into a bus-raping powerhouse, for less than a mortgage payment, then we'll talk.

If you live in London, UK, the average house price is around 800k USD. For less the mortgage payment (~6k) on that, I can get you a raid of SSDs (OK, not quite 256GB) and an extremelywilling bus.

Re:42 zillion dollars?

[mattcasters]

I actually RTFA and this device is obviously not remarkable because of the storage size. It's remarkable because of the transfer speed and the number of transactions/operations per second.
I can see how this device can be considered dirt cheap by people that are looking for a solution in that last area.

Re:42 zillion dollars?

[mrkitty]

SSD is fairly cheap on eBay

Re:42 zillion dollars?

[Anonymous Coward]

When this SSD is cheap enough that I can buy 3-4 of them and stripe that into a bus-raping powerhouse, [url]for less than a mortgage payment[/url], then we'll talk.

Dude, i just bought this house and the monthly mortgage payments are runing me 1.15 million a month @ 20%. I think you can afford your setup at slightly less than my mortgage payment.

Re:42 zillion dollars?

[Falconhell]

I am waiting for a SD card raid device, into which multiple say 8gig SDHC cards can be fitted. Such a device would be relatively cheap and easily expandable, and should perform great with RAID 5.

Re:42 zillion dollars?

[rainhill]

well cost/capacity goes hang in hand, one drives the other in the opposite direction, as in supply & demand.

Re:42 zillion dollars?

[Sabriel]

Probably not that far off. Addonics already makes a four-slot CF raid device as a PCI card, RAID 0/1/10. You'd have to use software for RAID 5 on it. I emailed them and they're looking at a PCI Express version. Can you get CF/SDHC adapters? http://www.addonics.com/products/flash_memory_reader/ad4cfprj.asp

This is good news...

[Tastecicles]

...if it can cope with high definition capture it'll be handy for me and my shutterbug family who're always out with various still and video cameras. Nothing worse than shortdropping a notebook and killing the hard disk.

How Much ?

[daveime]

For the price they currently want for these wizzbang solid state devides, I could buy a small data warehouse in Springfield, MA.

The price:

[amazeofdeath]

An arm and a leg, unofficial sources suggest.

Re:The price:

[tzot]

Well, that's a price drop alright. Until recently, you had to give an arm, a leg and your firstborn.

Finally

[Anonymous Coward]

Finally the speeds we are waiting for!

Large enough? No way.

[Yvan256]

A 256GB capacity is getting large enough to replace hard-drives for good.

Hard drives have entered terabytes territory, and you think 256GB SSD drives are "large enough"?

Until SSD drives cost only around 10-25% more than a regular drive of the same capacity, they're not replacing them at all. For most consumers, capacity is king, not speed.

Re:Large enough? No way.

[Aenoxi]

I fully agree with your conclusion that capacity is king for moist consumers, but... ...this is a 2.5 inch drive.

I'd like to subscribe to your reality if it has Terabyte-sized 2.5 inch drives. Where do I sign up?

Re:Large enough? No way.

[wamatt]

Remember the intended target market is the laptop crowd. 256GB is big enough to compete nicely. When it comes with those sort of performance figures, it's a no brainer if you have the money.

The current largest widely available 7200rpm is only 200GB. The majority of notebooks ship with 200GB of HD space.

Re:Large enough? No way.

[wamatt]

"Ship with LESS than 200GB space" I mean. slashdot removed my less-than-sign.

Re:Large enough? No way.

[Ant P.]

Let the warez kiddies have their 1TB disks then, this thing would be a perfect upgrade for my eee701.

Re:Large enough? No way.

[jht]

I dunno - I think once you hit that kind of capacity you can pretty much own the notebook market. Right now, mainstream notebook disk sizes are in the 160-250GB size range, with 320 generally available and I believe 500 GB drives are just starting to arrive. Most notebooks aren't at the high end of capacity, though.

I don't think SSD will make an impact in desktops anytime soon, but if I can put an SSD in my notebook and gain a little speed, some battery life, and better shock resistance without giving up any serious capacity (heck, my 2-month-old MacBook Pro has a 250GB HDD in it right now), depending on the price differential I'll probably be all over it.

Also worth thinking about (though it's not in the submitter's link) - I read a couple of releases on this drive yesterday, and though they aren't giving production prices yet they claim that multi-level cells will make it cheaper than the older models. Between that and the natural speed of price cuts, this drive may be at competitive HD pricing levels sooner than we expect. If I can get a 256GB SSD at a 25% price premium to a HDD of the same size (like you suggest), I think it would be pretty much a no-brainer. That 250GB HDD is only about $150 or so - maybe even less.

Re:Large enough? No way.

[amorsen]

For most consumers, capacity is king, not speed. For most consumers, price is king. Not price per gigabyte, just price. The only thing most consumers might need more than 32GB for is video -- and flash is almost as cheap as hard disks in the 32GB 2 1/2" segment.

Re:Large enough? No way.

[Bobby Onions]

Don't be so wet; there's no need to puch a dampener on things.

Re:Large enough? No way.

[Kjella]

2.5" HDDs are only in 500GB territory, that's only halfway there. If only prices would come down, they could easily stack a TB in 3.5" already. Prices aren't quite there yet but right now at newegg you got 80$/500GB = 0.63$/GB for regular HDDs and SUPER TALENT FTM20GK25H SSD 669$/120GB = 5.57$/GB so they're within one order of magnitude now. There's an incredible price development on SSD now and I figure they'll come close in a year or two.

Re:Large enough? No way.

[ozamosi]

I bought a new 120GB 2.5" disc last week - 256GB is "large enough" for now - if only the price was low enough...

Re:Large enough? No way.

[giorgist]

You can put your capacity in 50c blue ray disks (coming soon)there you go, you get your capacity and your speed

How long does it take you to fill 1 terabyte ?

(cue 740KB is enough for anybody meme)

So you have your 1/4Gig of super fast, super quiet, super low power consumption.
And every week or so you can burn a couple of blue ray disks for all your ... am ... storage needs

Re:Large enough? No way.

[aproposofwhat]

Try < - it works for me :o)

Re:Large enough? No way.

[Idaho]

Hard drives have entered terabytes territory, and you think 256GB SSD drives are "large enough"?

For notebooks? Absolutely. Even most new notebooks are not shipped with 250 GB HD's by default yet (although it is usually an option for a couple dozen extra bucks).

However, if it takes 2 years until this technology becomes mainstream, you may by that time well be right.

On the other hand, hard disk prices are really really low. By the time SSD is only about 2x more expensive I'd most definitely start considering it, since that would still mean paying "only" some $150. That should well be worth the performance increase as compared to, say, spending double the amount of $$'s on a faster CPU or what-have-you. Currently it's more like 15x as expensive though, so for now I'll stick with regular harddisks.

Re:Large enough? No way.

[thetoadwarrior]

I don't think the average person has terabytes of data. As long as this takes off for the lowest common denominator and they buy it up, then the prices will come down.

Re:Large enough? No way.

[cecil_turtle]

For most consumers, capacity is king, not speed. I don't know how you qualify "most consumers", but "sufficient capacity" is all most consumers need, after which price and speed absolutely do come into play. For most people 40GB is still sufficient capacity. Only people who download or rip very large amounts of music or movies need more than that, and that is pretty far from "most consumers" - your 14 year old son who tries to download every movie he's never heard of isn't "most consumers". Not to mention that IO speed is the performance limiting factor on pretty much every consumer PC nowadays, CPUs have been fast enough for a couple of years now.

SSDs and spinning disks can still co-exist - in a year or two you will be able to run your OS and programs on a 100GB-200GB SSD and go buy a 2TB disk or 5TB array to store your data on that is less performance critical.

Re:Large enough? No way.

[Yvan256]

Replying to myself here, about the comments made so far.

First, I didn't see it was for a 2.5" drive. As most of you pointed out, 256GB of high-speed, low-power storage in a laptop is a very good thing indeed.

To the few who say that most people don't have terabytes of data, you may be right. However a quick trip to your local Costco will show that external drives are now at 1TB and 2TB. If you want something smaller you will have to buy a 2.5" external drive which is around 250GB for nearly the price of the external 1TB 3.5" drive.

Re:Large enough? No way.

[FromellaSlob]

Let the warez kiddies have their 1TB disks then, this thing would be a perfect upgrade for my eee701. Have fun trying to put a 2.5" drive into an eee701.

Re:Large enough? No way.

[AbRASiON]

My OS partition is 25gb and that's me being extremely generous, I used to work with 7gb but thanks to XnView's thumbnail database and Picasa being a complete moron application without the ability to re-direct the cache database there too - I'm keeping 25gb to be sure.

256 is more than enough.

That being said, you wouldn't want to use Vista on this drive, what with the read / write constant disk thrashing it does because it's well ... stupid - it'll wear this sucker out in no time.

(It's sad that I'm old enough to remember a time where the drive light did not flash, EVER unless you did something, you could actually sometimes spot a machine with a virus just with un-authorised disk access back in the 6.22 days...)

Re:Large enough? No way.

[Anonymous Coward]

I disagree. My parents lived off a 15GB hard drive until a year or two ago when it gave up the ghost. They ran Windows XP Professional and had tons of space left over with no idea how little they really had. And from experience with other family members and friends, I suspect THAT is the norm.

I doubt the average user would spend a small fortune on a 256GB SSD, but they certainly would not fill it very fast if they had one.

For power users, gamers, and others who would use more than that, many will end up buying such SSDs anyways for the performance, reliability, and "coolness factor". That will certainly bring prices down to earth for the average user in time.

Re:Large enough? No way.

[Anonymous Coward]

in a year or two you will be able to run your OS and programs on a 100GB-200GB SSD and go buy a 2TB disk or 5TB array to store your data on that is less performance critical. This solution is ideal, since the real space-hogging files tend to be music or (especially) video, for which seek time is utterly unimportant ... once you've positioned the drive heads at the start of the content stream you just let the platters spin ... SSDs would be a waste in this scenario.

Re:Large enough? No way.

[mfnickster]

Agreed, GP needs a lesson in humidity.

Re:Large enough? No way.

[benhattman]

Mod parent up.

We are far past the point where the average consumer cares very much about capacity. What do you think they are going to do with 2 terabytes? Unless you are talking about someone who is frequently downloading movies and the like, I don't see how they would use that content. OK, there are probably a handful of people who are doing their own hi-def video editing or processing the output of large sensor arrays, but in what would do you define these guys as "most consumers?"

The reality is SSD doesn't have to come anywhere near the price of hard drives. It just needs to provide enough capacity (256-512 GB today) at a reasonable price. If you tell a consumer they can get a regular old hard drive, or pay 10% more for a SSD that doesn't fail when dropped and runs way faster, a lot of regular consumers will pony up for that.

Re:Large enough? No way.

[tab_b]

Not every machine needs to be storing lots of audio or video. If you can fit a Knoppix installation into a CD holding what, 640mb? 256gb is way overkill. If anything having desirable but smaller storage might discourage software bloat, leaving Linux and the *BSDs looking pretty good and leaving Vista in the dust.

Re:Large enough? No way.

[DogDude]

Capacity? Sure, 256 GB drives are down to $100 thee days, but do you know a lot of consumers that can fill a 256 GB drive? I've got my entire music collection from over 25 years of collecting music in FLAC format, and it's just now filling a 320 GB drive.

Re:Large enough? No way.

[Anonymous Coward]

I tell moist consumers that it's not the capacity that counts, but how you use it...

Re:Large enough? No way.

[DeusExCalamus]

You can put your capacity in 50c blue ray disks (coming soon)there you go, you get your capacity and your speed How long does it take you to fill 1 terabyte ? (cue 740KB is enough for anybody meme) So you have your 1/4Gig of super fast, super quiet, super low power consumption. And every week or so you can burn a couple of blue ray disks for all your ... am ... storage needs It's 640KB, actually....

Re:Large enough? No way.

[Anonymous Coward]

More than 32GB? Crysis is a 20GB game... 8 megapixel cameras with 2GB cards are becoming pretty standard... and Vista takes up over 10GB, throw office into that and your left with hardly anything to work with...

Re:Large enough? No way.

[TheRaven64]

My current laptop has a 160GB drive. May last one had an 80GB one, and that was only slightly too small. For portable use, I'd take a reliable 160GB SSD over a mechanical 500GB hard drive. At home I can always throw a few 750GB mechanical disks in RAID-Z for longer-term storage and not care if one of them fails.

Re:Large enough? No way.

[Anonymous Coward]

It's a well known fact that capacity is king for moist consumers.

Re:Large enough? No way.

[Firethorn]

Discounting microsoft bloat, when I look at stuff like this I think about people like my parents and grandparents.

You know, people who after having and using their machine for four years still has 80% of their 80GB HD free. Where the biggest increase in HD usage in the last year was microsoft patches.

For power users like me that DO get into games, video, and music on the computer, a 250GB SSD is enough to last quite a while. Heck, from initial build I'm likely to throw my OS and programs on the SSD and get the cheapest per GB HD(or two) for the rest of the stuff.

Going from 10GB to 100GB was 'Big', it enabled the start of consumer video. Going from 100GB to 1TB enables HD storage for the movies most people would watch in a month to a year. Current broadband speeds enables the downloading of HD streams in useful periods of time with a queuing system.

Basically, I'm saying that we've reached the point with HD storage that the majority of people don't need any more. They won't use what they already get on a bargain machine. It's like with CPUs. If you're not a power gamer*, the bargain basement machines will all run a cleaned up windows** and associated software with good speed. Or even one of the easier versions of linux and open office.

*My grandmother loved bejewelled. Mom does various solitares. They 'game' a lot, but their games aren't exactly demanding on computer systems.

Re:Large enough? No way.

[Firethorn]

Yes, but you could consider those external drives the equivalent of aftermarket turbochargers.

The vast majority of people don't need one, but those that do want the best they can afford.

In this case, the 'vast majority' of people who are getting them are the ones with enhanced storage needs. The second vast majority are the ones who simply get a 8-16GB USB stick and use that for their backup and sneakernet data transfer needs.

Plus, it's a lot more broken down that way - you only have around three factors to consider instead of the dozens for a full computer.

1. Storage capacity
2. Size (3.5" requires a wall wart, 2.5" doesn't)
3. Price

When it's that simple, many WILL figure out the price/GB, figuring if they end up buying a larger one for $10 more, but double the capacity, they're spending $10 as 'cheap insurance' against the unit being insufficient.

Re:Large enough? No way.

[pdusen]

Who the hell is going to install Crysis on a laptop?

Re:Large enough? No way.

[Anonymous Coward]

I am sure I am not a member of 'most consumers.' However, I have very little music on my system (some jazz and blues) and one tv/movie series (Firefly.) But as I am a game player I have over 200GB used right now. I could free up some space by clearing off games I have beaten, but I often go back and play through a portion of a game. 40GB just would not cut it for me.

Re:Large enough? No way.

[stonecypher]

Until SSD drives cost only around 10-25% more than a regular drive of the same capacity, they're not replacing them at all.

Mechanical random seek time is murder on certain business segments, particularly hosting, a vertical large enough that it made other seemingly cost-ineffective performance focussed storage technologies like SAN work. Between that and the seriously dropped cost of electricity, I think you'll find there's already a large viable market around the 3x cost line.

Considering that hosting has been buying solid state drives for years already, this one seems pretty clear cut.

Re:Large enough? No way.

[Bill, Shooter of Bul]

Apparently, you have no experience with high performance drives. 15k SAS drives max out around 300 GB.

See Cheetah capacities

SATA's and lower performing SAS drives come in higher capacities. I've never understood why. If some storage expert would like to explain I'd love to hear it.

Re:Large enough? No way.

[rainhill]

We are far past the point where the average consumer cares very much about capacity. What do you think they are going to do with 2 terabytes? You have no idea!

Re:Large enough? No way.

[Courageous]

After buying one of these things, it will be more like a lesson in liquidity.

Re:Large enough? No way.

[Anonymous Coward]

Well, I don't know about you, but I have at least 50 gigs of music plus photos I've collected over the years. I know plenty of other people like this, too. I collect about 5-10GB of stuff that is must-keep every year. So while in essence I guess I agree with you, it's not like most consumers don't have anything to store...

Re:Large enough? No way.

[Anonymous Coward]

Actually the new generation of 7200rpm laptop drives tops out at 320GB, ie Seagate 7200.3, Hitachi 7K320, etc. I totally agree with your point though - that was just an FYI :)

MLC, not SLC.

[Anonymous Coward]

But it's a Multi-Level Cell based Flash drive, not a Single-Level Cell based Flash drive. The cells hold 4 states, not 2.

High capacity, yes, and apparently high speed as well. Excellent... but also lower reliability. SLC Flash is extremely durable these days, but MLC Flash is not, last I checked, even one tenth as long-lasting.

How much lower? Well... ...frankly, we don't really know yet. We won't really know, as such, until they start to die - which could well be 5-10 years, and if so, that's really not bad - and you might not see the same type of can't-write-blocks failure, but rather a more conventional can't-read-blocks failure. Which would be about as bad as a hard disk crash (and we might have to develop whole new data recovery techniques).

Maybe it might last years longer than a hard drive owing to fewer moving parts. Perhaps it will slowly die, but good write levelling will largely mitigate the issue and overall it'll come out better, or about the same. Or perhaps we're looking at a flaky brick with lower reliability than a Quantum Fireball.

Early adopters, start your engines, because someone's gotta find out.

For enterprise use, it might be wiser to stick to more conservative SLC flash. Past that, all bets are off.

But we're seeing the beginning, here. Hard drives are, slowly, on the way out. It'll be a long phase-out where they are much more cost-effective for a long time... but it is coming. And I, for one, welcome our new nanosecond-seek-time overlords.

Re:MLC, not SLC.

[Anonymous Coward]

Normal SSD drives die in a matter of months on a typical developer's machine so it shouldn't be that hard to test.

I have first hand experience with this so I laugh when people say flash drives will last longer than their mechanical counterparts. The rewrite cycle count needs to be way, way higher than it currently is. Wear leveling can only do so much and it just gets worse as the drive gets full.

Re:MLC, not SLC.

[Anonymous Coward]

Post your on a site somewhere and submit to Slashdot then if you've uncovered this evil conspiracy of flash drive-life. Otherwise I'll just go ahead assuming that you're full of shit.

Re:MLC, not SLC.

[kipman725]

Quantum Fireball.. still got a 60mB one of those with loads of DOS games on that I think still works. Then agian it may be the exception that proves the rule.

Re:MLC, not SLC.

[Anonymous Coward]

Normal SSD drives die in a matter of months on a typical developer's machine [---] Care to back that up with statistics?

And I remind you that "I've seen it with my own eyes! And all my friends agree with me. And Random Dude on Some Forum agrees with me!" has zero statistical significance.

I have first hand experience with this so I laugh when people say flash drives will last longer than their mechanical counterparts. Anyone with a hint of common sense will ignore you because of the broad generalizations you're making, without providing any actual proof. I'm not saying that you're lying, just that you are probably confusing a broken unit with general behavior of SSD.

I have two SSD (SLC) drives. I don't trust them, because I simply don't know how reliable they are yet. So they are backed up slightly more regularly than my old mechanical drives. But until there arereliable numbers (Not "First hand experiences" from "Random Dude") I simply don't know.

Re:MLC, not SLC.

[flnca]

That's what RAID is for. Mirroring and/or striping with multiple drives and automatic error correction can make them more reliable. Bad areas of the SDD drive can automatically be excluded from storage by the filing system. That's basically the same thing with hard drives. A hard drive also accumulates bad sectors during its lifetime, and some day, fails completely. SSDs have a higher ruggedness when it comes to environment conditions.

Re:MLC, not SLC.

[Anonymous Coward]

We won't really know, as such, until they start to die - which could well be 5-10 years, and if so, that's really not bad - and you might not see the same type of can't-write-blocks failure, but rather a more conventional can't-read-blocks failure. Which would be about as bad as a hard disk crash (and we might have to develop whole new data recovery techniques). Don't worry; by then Macs will have taken over the world, and will be using ZFS as the native filesystem.

What? Why are you laughing?

Re:MLC, not SLC.

[multipartmixed]

I have a couple of systems that were built in 1999 -- no upgrades, ever -- which have been running continuously since the day they were built. The both have 12 GB fireballs.

Re:MLC, not SLC.

[Anonymous Coward]

I personally always did find talk about "wear leveling" insulting PR chit-chat. Yes, it really helps a bit, so it's not useless, but the improvement is quite insignificant if we are talking about 10,000 writes HDD. It's like bad joke for me. (yeah, I prefer when things work at least 4-5 years without a hitch)

Re:MLC, not SLC.

[alexo]

Spealing of SLC vs. MLC.

What are the best (fast, reliable) SLC-based USB keys?

Technology: Still new! Still Improving! Surprised?

[PC and Sony Fanboy]

Well, since the technology isn't developed, is it really that surprising that we read a story about 'Worlds Fastest' every couple weeks?

Solid State Drives for computers? They aren't really out of beta!

A 128GB SATA drive costs $3,000

[Poisonous Drool]

This drive could cost $5,000 based on a 128GB drive for $3,050 and 64GB drives from $900 to $1,150.

NewEgg search for Solid State SATA disks

Re:A 128GB SATA drive costs $3,000

[Anonymous Coward]

You are comparing to the wrong technology. SLC drives are extremely expensive, but this is a MLC drive. You should look at these items instead to extrapolate cost:

Super Talent MLC Solid State Disks on Newegg

The Samsung drive is much higher performance than these, but the fabrication costs should not be too far off. I'd guess $1500 for the 256 GB model when it comes out.

Re:Technology: Still new! Still Improving! Surpris

[Daniel Phillips]

Solid State Drives for computers? They aren't really out of beta! 400,000 eee PCs say you're wrong.

256gigs is a lot

[FudRucker]

I don't requires that much disk storage space, I could get by on 40 gigs and 80 would never run out of disk space for my purposes, make an 80 gig SSD that would sell for less than 200 USD and I will use my disk platters for target practice...

Re:256gigs is a lot

[flyingfsck]

I am getting by just fine on my Eee PC with a 4GB SSD and a 8GB SD card, for a total of 12GB.

The trick is to have a server somewhere with enough disk space to keep the bulk of the crap online.

Re:256gigs is a lot

[Anonymous Coward]

I'd be happy with an 8 or 16GB Sata300 drive to put my OS on. Everything else can be on the platters.

I don't need 32GB or even 64GB, I need enough space for an OS, no more no less. Everything over that is pointless. (until the Terabyte flash drives come out).

Re:256gigs is a lot

[Idaho]

I don't requires that much disk storage space, I could get by on 40 gigs and 80 would never run out of disk space for my purposes, make an 80 gig SSD that would sell for less than 200 USD and I will use my disk platters for target practice...

Exactly, my current notebook (3 years old) only has a 60 GB disk, and even then I have split it into 2x30 GB so I can run both Linux and Windows XP. I've never had a problem related to lack of space. OK, you can't keep huge movie or music collections on it, but seeing as how laptop sound quality is typically rock bottom anyway (the soundcard as well as built-in speakers), I don't normally use it to play music.

30 GB is apparently plenty to store whatever software I need for work (which is quite a range of applications) as well as some games and movies in case I'm traveling for a couple of days.

Clearly I'm not using it as my "only" computer though. In that case 60 GB definitely would not suffice, by far.

Re:256gigs is a lot

[v1]

just thinking about it, although having a fast reboot time is surely nice, it's not something you should have to benefit from often? I beat my system to death here and reboot about every other week. I know some people that haven't restarted their machine more than a dozen times since they bought it a year ago.

Re:256gigs is a lot

[Anonymous Coward]

If you eliminated the extra your extra s's, you'd need even less space.

Re:256gigs is a lot

[Bryan Ischo]

I am so totally with you. I have never used more than about 20 GB of any workstation hard drive I have ever owned.

But, I don't steal videos, music, and software off the internet, so I guess I don't represent the majority of the market.

Re:256gigs is a lot

[Jeff DeMaagd]

But, I don't steal videos, music, and software off the internet

I don't either, and I still have terabytes of data, such as audio and video files.

Depending on how it's encoded, 1 hour of video can take 1GB to 13GB without even going to a professional codec. ATSC HDTV recorded from an antenna easily takes 6GB per hour. My AVCHD camcorder runs at about 8GB per hour, converting it to a more editable intermediate codec easily runs 13GB an hour.

Re:256gigs is a lot

[ABasketOfPups]

Lossless codecs like huffyuv run about twice that hot for SD res, around 18-30 gigs/hour (depending on how clean your source material is...). Heaven help you if you're doing lossless HD... or lossless, high-color HD... might as well go buy stock in hard drive company now.

Re:256gigs is a lot

[Jeff DeMaagd]

Heaven help you if you're doing lossless HD... or lossless, high-color HD... might as well go buy stock in hard drive company now.

Isn't that the truth. Right now, Redcode RAW goes up to 130 GB/hr. Despite the name, it is actually lossy. I guess lossless 4k out of the Red One camera might be about 1TB/hr.

Re:256gigs is a lot

[Bryan Ischo]

Yeah after I posted that I knew someone would respond that digitally recorded home/professional video takes tons of space too. But I think that isn't very mainstream either.

Re:256gigs is a lot

[mcrbids]

Wow. You think 256 GB is alot? I have that on my laptop. I have almost 2TB of total storage on my home network, despite me tending to delay upgrades until the last possible second. (Some 20 GB drives, there...)

No, it's not all MP3s and MPEGs. In fact, that's a very small minority. Mostly, it's word files, email, VMWare sessions, and a few ISO images. It's just stunning how rapidly these things add up when you use them everyday at work!

The problem that SSD drives face is something like the "two guys running from the bear" joke, which goes something like this.

Two guys are hiking thru the woods, and they come across a bear. Both guys start running as fast as they can. One guy says to the other: "Do you think you can outrun that bear?" To which, the other guy says: "I don't have to. I only have to outrun YOU."

Flash drives have made excellent progress, improving FASTER than hard disk technology has advanced (which has been significantly faster than Moore's law) but it still faces (as it always has) the challenges that a slightly older, more mature technology manages to perform somewhat better in that all-important arena, the cost/per/unit arena.

And because of that, flash/SSD will only do better in those areas that flash is better positioned against HDDS: low churn, low power, compact storage. (EG: MP3 players, ipods, phones, etc)

Yeah, they'll be 1/10th the price in a few years. So will hard drives. And so, against the bear of economics, HDDs outperform SSDs and remains ahead.

Re:256gigs is a lot

[Explo]

How about digital photography? It's a fairly common hobby these days and a single jpeg from pretty much any camera, whether a P&S compact or a DSLR has a size of several megabytes. It doesn't require hardcore dedication to snap thousands or even tens of thousands of photos in a few years. My collection of stills dating back from 2002 takes currently up about 50 gigabytes. While I might be more active photographer than average, filling 20 gigabytes worth of storage with photos isn't really very hard, if the user isn't actively removing not-so-important photos every now and then.

(I also have several gigabytes of completely legal music ripped from my own collection of CDs on the HD; I'm too lazy to rip everything, but it's handy to keep some of my favourites permanently available with minimal effort)

The Power of 1000 Hard Drives

[StCredZero]

This pales in comparison to the ioFusion drive. The videos show tests being run where they are doing 8 operations at the same time, at blazing speeds, copying multiple DVDs in 5 seconds, and simulating swapping a blizzard of 4kb blocks as fast as RAM. Instead of 2 channels, their cards use 160 channels at the same time. This gives a single card the parallel random access bandwidth of a 1000 disk drive SAN.

http://www.tgdaily.com/content/view/34065/135/

At $30 per gigabyte, it would be great to have a 10-gig for OS and your current favorite MMO game.

Re:The Power of 1000 Hard Drives

[Anonymous Coward]

"At $30 per gigabyte, it would be great to have a 10-gig for OS and your current favorite MMO game." Why? Running Vista with 4GB ram my HD isn't accessed after it's been booted. Can raid all night long without my HD being any kind of bottleneck.

Re:The Power of 1000 Hard Drives

[StCredZero]

Eve Online had big problems with random-access hard-drive use with the local cache. I haven't played in awhile, so I don't know if this is still the case.

Re:The Power of 1000 Hard Drives

[custompccases]

"At $30 per gigabyte, it would be great to have a 10-gig for OS and your current favorite MMO game."

Unless you play Age of Conan, roughly $1050. That would pay for like 2 months of subscription and the cheetos required to sustain the individual.

Re:The Power of 1000 Hard Drives

[Anonymous Coward]

This is fscking crazy - insane speed increase ! I just researched a bit on the net, and it looks like they've already started deploying (according to the website, they're deploying to 44% of Fortune 100).

Still a long way off from consumer technology, since they seem to be targeting enterprises right now. I know I'd want one of these on my database boxes.

*Easily* large enough.

[Whiney Mac Fanboy]

Hard drives have entered terabytes territory, and you think 256GB SSD drives are "large enough"?

Yup, even 64GB is easily large enough for my primary hdd on my laptop - I'd pay a premium far larger than 30% if the price/performance relationship was linear.

Until SSD drives cost only around 10-25% more than a regular drive of the same capacity, they're not replacing them at all. For most consumers, capacity is king, not speed.

But, the bulk of my content resides on network servers (same holds true for my less geeky friends - but for them, substitute 'usb drive' for 'network servers').

Primary SSD + Storage HDD = Gold

[SD-Arcadia]

Once the prices come down and the tech matures a little more, a nice small 32-64GB SSD for the apps and a 1TB+ for storage should be a great overall solution. This could even happen in form of an elegant hybrid unit.

Did we not already have?

[nnxion]

Large and very expensive SSDs?
http://www.gadgettastic.com/2007/10/05/fusion-io-launches-the-iodrive-640gb-pcie-hard-drive/

$19,200 for a 640 GB Hard Drive...is there even a market for such things now? Consumers are already used to having hard drives those sizes, hopefully it won't be too long until they find a good way of pushing those prices down and making them more accessible to the general public.

Re:Did we not already have?

[PopeRatzo]

$19,200 for a 640 GB Hard Drive

We all know that's more expensive than computer storage has ever been before.

Right?

And if hard disk storage had ever been that expensive, it would have meant the abandonment of the hard disk technology forever.

Right?

Re:Did we not already have?

[c6gunner]

We all know that's more expensive than computer storage has ever been before.

Right?

And if hard disk storage had ever been that expensive, it would have meant the abandonment of the hard disk technology forever.

Right?

THat's a specious argument. It doesn't matter what HD's cost 30 years ago, what matter is how much they cost now. And if you don't agree with that, then I'm sure you'd be happy to buy my vintage late 80's 20 megabyte HD for $600, right?

Re:Did we not already have?

[PopeRatzo]

I'm sure you'd be happy to buy my vintage late 80's 20 megabyte HD for $600, right?

No, and that's the point. Just because SSD drives are extremely expensive now does not mean for one minute that there is some fundamental flaw that will prevent them from overtaking hard disks as ubiquitous computer storage. In fact, if history is any teacher, The machine I build from parts bought on newegg or tiger direct in 2010 will have SSD components, even if only the system drive.

As a creator and purveyor of digitally-made music, I can tell you that SSD technology sounds like something that will make an improvement in my work and life. Just never having to hear the "clunk-clunk" of a hard drive getting ready to shit the bed will be a great improvement.

Re:Did we not already have?

[Anonymous Coward]

Your original point was flawed. At the time that hard disks cost tens of thousands of dollars, there was no competing technology to choose from. So companies would by $10,000 1 gig hard drives because it was the only way to get what a hard drive could offer.

Now with SSD drives, there is competing technology. Sure, SSD drives have many features which are nicer than hard drives, but in the vast majority of applications, these extra features do not justify the cost; they don't even come close.

I do believe that SSD drives are the better technology and that they will win in the long run. This does not make your original point relevant however.

Re:Did we not already have?

[rasputin465]

$19,200 for a 640 GB Hard Drive... Well, they have to be able to pay the models who pose with the product.

Re:Did we not already have?

[billcopc]

False argument.

Back then, there was no alternative. Today, the common sysadmin can cobble together a combination of striped disks and aggressive memory caching that either matches or beats SSD performance for a small fraction of the cost. The gap between disks and SSD is not large enough to justify the added cost, except in very exotic scenarios.

Should we abandon SSD research ? No, I don't think so, but I do think they need to stop advertising every little bunny-hop until they have something truly revolutionary in their hands. Wake me when they can saturate a 300mb SATA port with a single SSD, or they can run 10x more IO/s than the fastest 15k disk.

Re:Did we not already have?

[Anonymous Coward]

I'm quite sure that the 2.5 Mb 12 inch disks that came with PDP 11-s where _much_ more expensive per megabyte.

That is, they did cost much more than 7.5 cents (19200 / 640000 * 2.5) apiece

Re:Did we not already have?

[Anonymous Coward]

"We all know that's more expensive than computer storage has ever been before."

Being young, aren't you? ;)

Re:Did we not already have?

[Anonymous Coward]

>> $19,200 for a 640 GB Hard Drive

>We all know that's more expensive than computer storage has ever been before.

>Right?

Wrong.
How is this "insightful" in any way? It's actually totally clueless - to value storage only based on capacity.
$19.2K of SSD storage is probably already cheaper than traditional HDD (measured in IOPS).

In practice SSDs are still slow as hell

[Britz]

Except the ones that cost more than four figures. So don't try affordable SSDs just yet. You would be very disappointed.

Time to redesign the personal computer?

[yancey]

With high capacity non-volatile memory, is it now time to redesign "personal computer" hardware and the operating systems?

Re:Time to redesign the personal computer?

[FurtiveGlancer]

Definitely. How much faster could life be if every program were permanently resident in memory? Buh bye pagefile, hello performance.

Re:Time to redesign the personal computer?

[symbolset]

T minus 8 Days.

Re:Time to redesign the personal computer?

[4D6963]

A most interesting and pertinent question! I think that if such a memory reached the speed of RAM with the capacity of a HDD then we could merge the two concepts into a central memory that would be used for anything. The first real gain with that type of design is that instead of loading (uncompressed) files (from the HDD to the RAM) you could simply point to them, and directly access them. Virtual machines could benefit greatly from that by pausing and resuming their execution instantly, for all their virtualised RAM would be written in a file that would simply pointed to. The same could happen for regular programs. They could have all their memory space in a file (the OS would take care of it), and if the program was to be prematurely killed you could resume its execution state.

Likewise, it would remove the concept for RAM space, as well as for virtual memory, that is, the OS wouldn't use a single file to put everything in, but rather as many files as it needs for each program (for example). With such a concept the execution for everything I mentioned (programs, OS, virtual machines) could be paused and resumed instantly.

As for the actual booting of the machine, I'm sure a clever use of it by say copying read-only pieces of memory that are hardware/configuration-independent to another space in memory where they could be modified (or not, maybe you could have a partially read-only OS) would greatly speed up things.

Somehow I can see that happening in embedded devices, not so soon with desktop machines, but we'd have yet to wait for SSD memory to be fast enough.

Re:Time to redesign the personal computer?

[illumin8]

A most interesting and pertinent question! I think that if such a memory reached the speed of RAM with the capacity of a HDD then we could merge the two concepts into a central memory that would be used for anything.

Interestingly enough, Solaris had this concept since at least the early to mid 1990s. I remember I could allocate a chunk of RAM to serve as a "caching" RAM filesystem in front of any back end storage device. Any local disk, NFS server, CD/DVD disk, or mountable filesystem could have a 500GB or so caching RAM filesystem in front of it. The data would be read from disk once, then accessed from nearline memory at all times.

At that point, your outdated spinning media (aka hard disk platters) actually become a "near-realtime" backup system, where cached writes are eventually written out to disk as soon as the kernel has the spare cycles to flush the write cache.

I truly think storage is moving more towards this type of system with fast RAM, slower RAM, and disk. Operating systems will have to adjust their outdated concepts of "free memory" and just have "free fast memory" and "free slow memory" and intelligent processes that pre-fetch and copy data between the slow and fast zones in anticipation of which applications and files the user accesses the most. Hell, if Solaris could do it 15 years ago, I'm sure Mac OS X or Linux could do it today.

DBAs: Index tablespaces? Logfiles?

[Jeppe Salvesen]

How would this perform for index tablespaces and logfiles? I imagine lifetime/health will have to be monitored, but that's already being done with regular platterspinners.

Re:DBAs: Index tablespaces? Logfiles?

[KokorHekkus]

If you look at the lifespan of an SSD just using the limited write-cycle aspect is a function of how many physical blocks you have * number of write-cycles (if you have a proper wearlevelling algorithm that swaps writes from most used block to the least used ones).

From what I've read physical block size usually are in the range 16 kbyte - 256 kbyte. Let's assume a low 100000 cycles, a highish 256 kilobyte blocksize on a 256 GB drive giving you 1 million physical blocks. If you use that to estimate the number of writes/second you can use 24/7 over a 10 year period you get 10^10 writes per year which gives somewhat over 300 writes/second. Sounds acceptable to me and I'm guessing my estimate is on the conservative side.

Re:DBAs: Index tablespaces? Logfiles?

[sshir]

You don't really need it for logfiles - sequential writing is fast enough on non-SSD drives.

Re:DBAs: Index tablespaces? Logfiles?

[DDumitru]

Most current Flash SSDs have a 2 Megabyte erase block. Take a 32 GB MLC drive with 10,000 cycles and you wear out in 160 million random write operations. At 20/sec this is 92 days. At 200/sec it is 9 days.

Good enough for laptops. Not good enough for servers.

but ...

[jsnipy]

But does it blend?

Re:Early adopters, start your engines

[v1]

Early adopters, start your engines, because someone's gotta find out.

I was rather expecting a "iWantOne" tag on this article, because I DO.

I've been an early adopter on hard drives more than once. Back in '98 my laptop had a 23gb (yes, 23) HDD in it, and that was awesome to have that kind of portable storage. It made that nasty "I'm about to die" click about ten times a day, for every day of the two years I owned it too, when I sold it in working condition.

If it's not too painful I may bite. My laptop syncs with the backup every night so I'm not too worried if it tanks. In fact I hope it does. That means they'll give me a new, better one a couple months later. Maybe more than once. That can be one advantage of early adopter. By the time things settle down, you have the same thing that everyone else did, but you've just had it a lot longer. (at a price of course)

Too bad TFA didn't give a guesstimate on MSRP. One thread I found suggests $8k which is a little steep even for me.

One person tested a MBP booting off a 64GB SDD. OMG. The gear didn't even have a chance to spin. A good chunk of the boot time was taken waiting for hardware to come ready.

256GB is more than enough!

[Anonymous Coward]

I'm thinking of time critical operations where the HDD's access times create a bottleneck such as running an operating system or programs with many simultaneous file accesses. Storing huge chunks of data (music/movies) will still be the domain for (probably external) HDDs for some time, but the primary boot disk with the applications (which one most systems amount to ar less than 256GB) will shift towards SSDs once the prices come down. It's just the logical step to maximize performance.

Is there a good reason why....

[oztemprom]

this flash type memory is not being incorporated into existing platter based HDDs? It would seem to me that a few (read 8 - 64) GBs of flash memory coupled to a .5 to 1 TB standard HDD would be a great easy to use product.

Re:Is there a good reason why....

[Anonymous Coward]

you would be increasing complexity of the whole thing, making it more prone to failure and having the flash mem quasi as a cache for the hdd would help you with the speed for writing, but not so much with reading. also you forego the advantages of a pure SSD that are: no moving mechanical parts, less energyuse and heat and noiselessness.

Re:Is there a good reason why....

[menace3society]

I think the problem is, if you try to use the flash as a sort of buffer, you end up with the worst of both worlds. You're still subject to the same mechanical failure risks for long-term storage as a simple hard disk, but you're also limited by the number of writes you can do to the buffer.

You can accomplish the same thing, with fewer flaws, by just having two drives.

Re:Technology: Still new! Still Improving! Surpris

[PC and Sony Fanboy]

I guess you're right ... if we're talking about 8 gig SSDs in tiny, underpowered but portable netbooks. SSDs have been in use in cellphones for awhile .... as flash memory. The EEE is only slightly more functional than a full featured palm device or blackberry, and both use flash memory as well.

I was talking about home computers, work computers, computers that do more than just check email and the occasional word processing... I've got a solid state 'drive' as a flash drive right now bigger than the 'hard drive' on some of the entry level EEEs ... I certainly don't think it would be sufficient for a full function computer...

Gamers will love it

[Anonymous Coward]

This will be great for gamers to use as a primiary drive and then use more safer HDD's as storage. Can't wait to see the prices

Am I the only one who is fine with 80gb???

[distantbody]

...and you think 256GB SSD drives are "large enough"?

For me, 80GB leaves plenty of room after apps,data, docs and games. Am I using my internet wrong?

Re:Am I the only one who is fine with 80gb???

[SelrahCharleS]

I must be using it wrong too. I dual boot XP and Ubuntu on a 60GB hard drive. I've started filling it up, but that could be remedied by deleting the different Linux ISOs I left sitting on my documents partition.

Re:Am I the only one who is fine with 80gb???

[Anonymous Coward]

Am I using my internet wrong?

Yes, because you apparently do not have a pr0n collection.

Re:Technology: Still new! Still Improving! Surpris

[Anonymous Coward]

Just because a specific technology is being sold to the public is NO indicator that technology is out of beta. See:

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

and please try to keep up. Thanks.

Apples to Celery

[benhattman]

You can't just compare different markets. As another poster said, you can buy CD-R for less than a penny each. What you are referring to is how record companies have used the lower medium price to make an even larger profit off of the content.

However, how does an oligopoly selling copyrighted content compare to a commodity market? Basic economics tells you they don't, and you can count on one of two things happening. A) SSD prices fall in line with hard drives. Or B) hard drive capacity moves beyond the needs of most consumers and SSD takes up that niche while being only marginally more expensive per GB than hard drives.

Why so much?

[Anonymous Coward]

Maybe this has a use for enterprise, but I have a theory that if you gave a normal home user 25 GB of solid state and used it as a cache for 250 GB of traditional HD, you'd get nearly the same performance as 250 GB of cache--think about what volume of data you access on a given day, how rapidly that "working set" changes (I estimate 2GB a day for myself), and note that that change is all the disk IO that would be necessary.

Re:Technology: Still new! Still Improving! Surpris

[benhattman]

Solid State Drives for computers? They aren't really out of beta! What would you call an iPhone or iPod Touch? 10 years ago, these are top of the line computers in your pocket. Of course, they are a little crippled by Apple, but the point is SSD is in all sorts of real computers. Unless you only count it as a computer if it has a keyboard attached.

A little off

[tab_b]

Actually $80/500GB = $0.16/GB, so the gap is even wider between the Super Talent's $5.57/GB and say a Seagate 7200.10. Right now the price per gigabyte of SSD vs traditional is almost 35x higher, but it'll certainly come down drastically in the next few years. When it gets down to under 5x higher, I think it'll really take off.

Re:Technology: Still new! Still Improving! Surpris

[PC and Sony Fanboy]

you should try reading the comment thread (similar to reading the article) before posting. Really, it would help you avoid looking stupid.

What kind of filesystem?

[Anonymous Coward]

That type of functionality can be put into a filesystem.

Is this raw access, or over a filesystem? If it's the former, you have a benchmark which doesn't mean much in the real world. If the latter, which filesystem was used?

Choosing the wrong filesystem type will indeed get you non-optimal performance.

The serializing of the random read-writes

[crovira]

is the job of the disk controller which takes the read (or write) request sorts the requests by the sector address, takes the current position and direction of the head positioning mechanism and performs the requests in the physical sequence.

You may request something but other requests may get done first because the heads are positioned better.

IBM made their money by doing that and protected it with measures that skirted with antitrust violations throughout the sixties and seventies.

Its a kinder and gentler way of doing it and it was originally done to preserve the hardware and stop needless thrashing. That it ended up being more efficient for multi-tasking/multi-processing computers was just icing on the cake.

Re:The serializing of the random read-writes

[hairyfeet]

Which is why I have been wondering: You know how a HDD will have a bit of RAM as a cache to make write more efficient,why aren't they doing that for SSDs? It seems like it would me a lot more efficient if lets say,the SSD has 4Mb blocks to have 8Mb of cache on them so the data can be cached and written in efficient blocks. I'll admit I haven't gotten to read up on the algorithms used with SSD so I don't know if there is a technical problem with that idea or not,but it would seem to me to be a "best of both worlds" kind of solution.

And speaking of "best of both worlds" what happened to everything is going to be hybrid? A couple of years back all you read was they were going to add anywhere from 256Mb to 8Gb in addition to the 8-16Mb of RAM cache on HDDs to boost the data access and make them even more efficient at writes. What happened? I know I would personally like a hybrid that had,say 8Gb on it so I could have the OS stored in flash with my data and swap stored on the platters. But that is my take on it anyway,YMMV

Re:The serializing of the random read-writes

[Wesley Felter]

High-end SSDs have a RAM write buffer, but it complicates the controller significantly. I expect sophisticated controllers to trickle down to all SSDs eventually.

Re:The serializing of the random read-writes

[Waccoon]

I'd like to see an 8GB flash drive piggybacked onto a 5400 RPM hard drive. Performance is hardly relevant for the 60GB of videos you have on your hard drive, and it would be especially nice if you could partition the flash and hard drive separately. That would really offer a cheaper, quieter, smaller solution for a laptop without skimping on space.

Are there any OSes that natively prioritize frequently-used files on fast partitions? I understand that mixing fast and slow memory on a motherboard isn't cheap due to requiring multiple memory controllers, but mixing slow and fast hard drives makes a heck of a lot of sense to me compared to spending tons of cash on striped RAID.

Re:Random read ops?

[aharth]

How about random reads? I've benchmarked a 16G Samsung SSD and the standard Linux file systems (ext2, ext3) seem to cache read blocks in the (main memory) file system buffers.

Doing so seems to diminish some of the the possible overall system performance improvements - if I have a SSD I want to use the main memory for either HD io caching or programs. Caching disk blocks from the fast SSD in main memory seems suboptimal.

Capacity is just one factor

[Anonymous Coward]

A 250GB SSD is a great, but many people are doing just fine with less. What doesn't seem to have been mentioned so far is the resulting increase in battery life, the reduced heat output or lower chance of drive failure one would get with an SSD drive. Those for me are much more important than the drive capacity.

What about memory wear?

[popra]

to my knowledge, SSDs use flash memory which used to (?) have the limitation that it "wears" out on write-erase (http://en.wikipedia.org/wiki/Flash_memory#Memory_wear).

Was this limitation removed? Or are there improvements that make the limitation irrelevant? How would a SSD compare from this point of view to a standard HDD?

Re:What about memory wear?

[jjohnson]

Wear-leveling algorithms distribute the write-erase load relatively evenly across the SSD. Individual NANDs have a MTBF of around a million write-erase cycles, IIRC, so good leveling algorithms have already pushed the MTBF for the SSD past the expected lifespan of mechanical HDs.

I like SSD but....

[jeffc128ca]

I've been waiting for something to get around the hard drive speed bottle neck for a long time. I do a lot of data analysis on huge data sets, mostly financial market data. I end up doing a massive amount of reads and writes to hard drives which slows things down a lot.

My main fear with SSD's is the wearing out of blocks and bits. Typical data sets I work with are about 2 gigabytes. I run scripts against the data to look at various patterns and generate forecasting data. I could read and write that data six or eight hundred times in a day's testing. Well over a terabyte a day. How soon before an SSD craps out on me at that pace?

I would love to have an SSD for the blazing fast access times, but I don't want to have to replace it every six months. I'd pay extra for it, probably 2 to 3 times the traditional hard disk amount. But it has to last a few years at least. The other option of going 64 bit, adding huge amounts of DRAM, and running a RAM disk isn't financially sound at the moment.

Re:I like SSD but....

[toddestan]

Just curious, if you're main problem is the hard drive speed, but your datasets are only ~2GB in size, why don't you just build a machine with a massive amount of ram? I figure you could easily have a computer with 16-32GB+ of DDR2 ram for a few grand. Seems like a cheaper and faster way of dealing with it instead of moving to flash disks.

Re:I like SSD but....

[Courageous]

Anyway, for single cell flash, I would think that at your volume you'd be obsoleting the drive before it wore out. For multi-cell, I'm not so sure.

If you're working in the financial services industry, shouldn't cost not be that much of a factor?

C//

Re:I like SSD but....

[Anonymous Coward]

We have 16 gigs of ram in one of our servers. Keeping our dataset all in ram is a dream come true.

Too bad it's not financially sound for you.

Re:I like SSD but....

[jeffc128ca]

At work I have the luxery of expensive server equipment to play with. It's my home projects where I have a budget that is limited. However I am getting closer and closer to what your talking about. An 8gig ram drive would be perfect and the cost of getting that much ram, and the motherboard to hold it, is getting closer to a my budget. Most consumer stuff tops out at 2 or 4 gig for memory at the moment.

It won't be long from now when consumer home built rigs will start to have some serious data center level muscle. Extra memory and faster disk access go a long way to getting there.

Re:I like SSD but....

[rdebath]

Maybe if you only shop at Dell, but I've just bought a machine that seems to fit your requirements. It'd probably be less than a thousand dollars.

1 x Intel Core 2 Quad Q6600 processor
1 x Gigabyte GA-P35-DS4 motherboard
2 x OCZ 4GB Kit PC2-6400 DDR2 (Gold Heatspreader)
1 x PNY GeForce 7 7200GS 128Mb PCI-E Graphics Card
1 x LITEON 20X DVDRW BLACK BARE INC SOFTWARE
2 x Seagate Barracuda 1TB ES.2 SATA 3Gb/s Hard Drive
1 x Antec nine hundred EU games case
1 x Ultra quiet Active PFC 500W PSU with blue LED fan

Have you noticed how everything seems to have blue leds these days!!

You need to chop off a little bit of the start of those terabyte disks to get you fast filesystem area, the rest can be raided archive. There are lots of drive bays and sata ports for whatever other flash or hard drives you want. You may want a better graphics card. If you app is single threaded you'd be better off with a dual core for the same cost.

Wear leveling

[tepples]

And since it's using MLCs, expect to buy another one quite soon after. While SLC flash is up to the 100,000 to 1,000,000 rewrites level, MLC is still closer to 10,000. This, combined with the larger cell sizes on most MLC products means that it is likely to wear out much faster. MLC products use wear leveling at the controller level to spread writes to high-traffic areas such as directories and extent maps around the physical medium. That's also why they store 256 GB and not 256 GiB: that's 7 percent for spare sectors. Even if this wear leveing is only 10 percent effective, how long does it take to write to all 256 GB of the drive 1,000 times?

Toshiba SSD...

[barl0w2]

Some SSD love from Toshiba from CES 2008: http://flickr.com/photos/barl0w/2179248913/in/set-72157603667187312/

BAARF

[tepples]

Easier to miniaturize, certainly. Right now they're doing massive RAID0 to get performance, I wonder what it'd be like if they could do RAID1/5/6 for example - forget hard disk crashes more or less, just replace some flash plug-in modules in your SSD. Ok the electronics could still fry, it could get lost or stolen but mechanical failure seems to be the typical killer. I've read that RAID 3/4/5 is unreliable. As capacities grow, it takes longer to reconstruct a new spare from the surviving drives when one dies. In fact, BAARF contends that capacities have grown to the point that it's likely that another drive will fail during reconstruction. Are there any big drawbacks to RAID 6?

Re:BAARF

[WuphonsReach]

I've read that RAID 3/4/5 is unreliable. As capacities grow, it takes longer to reconstruct a new spare from the surviving drives when one dies. In fact, BAARF contends that capacities have grown to the point that it's likely that another drive will fail during reconstruction. Are there any big drawbacks to RAID 6?

RAID6 is a far better option then RAID5. At least it makes it less likely that you'll end up with a double-drive failure that takes out the entire array.

OTOH, the failure mode of both RAID5 and RAID6 leaves a lot to be desired. Rebuild time increases linearly as you add more disks to the array. So a 10+ RAID5/RAID6 array can have huge rebuild times, leaving you vulnerable for a lot longer. As in half a day or longer to rebuild the array (or at least a few hours).

Personally, my preference is the more conservative RAID10 approach. Rebuild times are based on the size of an individual disk in the array (not the total array size), which means your vulnerability window is a lot smaller. And depending on luck, you can survive a multi-disk failure. Rebuild times are typically under 2 hours for arrays that are based on 300-500GB drives.

(My preference is to have 1 spare disk for every 6-8 drives in the array. So a 12 disk RAID10 array would probably be RAID10 over 10 disks with the other two as spares.)

Patents.

[tepples]

However, how does an oligopoly selling copyrighted content compare to a commodity market? Recent patent lawsuits by SanDisk and Seagate show that the market for flash memory is closer to an oligopoly selling content subject to exclusive rights than to a commodity market.

To work around power management defects

[tepples]

just thinking about it, although having a fast reboot time is surely nice, it's not something you should have to benefit from often? At least on a laptop, a faster boot would help compensate for cheap and/or buggy hardware that doesn't come back online properly after a sleep.

Revision control?

[tepples]

A most interesting and pertinent question! I think that if such a memory reached the speed of RAM with the capacity of a HDD then we could merge the two concepts into a central memory that would be used for anything. Newton OS and Windows Mobile 4 already did this.

They could have all their memory space in a file (the OS would take care of it), and if the program was to be prematurely killed you could resume its execution state. How would revision control work under your system? The naive revision control supported by most document-editing applications has three revisions: a "current version" loaded into working memory, an "undo buffer" also in working memory, and a "stable version" saved to a persistent file system. "Save" copies the current version to stable; "revert" does the opposite. The Newton OS's unified memory model had problems with rollback: once you made two unwanted changes to a document in NewtonWorks, you couldn't recover the document because there was no stable version.

Re:Revision control?

[4D6963]

How would revision control work under your system?

The same way as it works right now, that is it's handled by the program? Under my system, things work normally, just instead of having a space in RAM and the paging file dedicated to the process, you may have a file dedicated to the process' memory on the file system. Note that you would point to a file without loading it/copying it presumably only in read only. Or did I misunderstand the question?

Making a YouTube video

[tepples]

Only people who download or rip very large amounts of music or movies need more than that, and that is pretty far from "most consumers" But nowadays, with Web 2.0 and user-generated content and all that, are most PC users "consumers", or are they also "producers"? How much storage do the 100 minutes of raw footage edited into a 10-minute SDTV video take?

Video editing

[tepples]

For me, 80GB leaves plenty of room after apps,data, docs and games. Am I using my internet wrong? It appears you might not be using this part of the Internet. How much space does an hour of SDTV footage that gets edited down into a six-minute video take?

Re:Random read ops?

[Wesley Felter]

So use O_DIRECT like some databases do.

New product from Dell......

[Rooked_One]

So, Dell is about to offer up their MD1120 - same exact thing as the MD1000 but with 24 drives - 2.5" each.

If you didn't care about redundancy, for a DB this thing would be perfect... Mix and match slower sas drives in an array that doesn't require such fast IO, then on your data intensive VD, have SSD's!

That would be pretty slick.

Re:Random read ops?

[Daniel Phillips]

if I have a SSD I want to use the main memory for either HD io caching or programs. Caching disk blocks from the fast SSD in main memory seems suboptimal Bandwidth to the SSD is 300 MB/sec (sata 2) while bandwidth to main memory is more than 10 GB/sec, 33 times faster.

SDD and applications

[owlstead]

I do hope that operating system and application programmers will be wise and start to program to cater for faster system drives and slower storage drives. It's a shame if a game uses a real fast drive just to load some levels. Safe the big number data on a "slow" moving drive and store the engine and some much used textures on the main drive.

256 GB SDD (or a spanking new 300 GB Velociraptor for that matter) can be lots of fun, but not if each and every game stores 40 GB of uncompressed game data on the drive. This could easily fit on a single layer, single sided BD disk, so this cannot be far off.

SSD on the desktop

[Anonymous Coward]

The largest issue with any SSD is the fact that most desktop class OS's are not tuned for them. XP in particular is hard coded to expect a ~10-12ms access time. I don't claim to be a guru on the IO habits of MS OS's but most modern OS's have implemented many kernel code level 'gymnastics' to avoid hard disk latency that is simply not an issue with SSD. When these algorithms are presented with low latency drives, such as a SSD, they don't compensate for the added performance.

Example:
The Linux elevator CFQ IO algorithm will sweep the disk, perform all requests when the head passes the required location. IO's are investigated and re-ordered based upon several factors, the largest being the head location. With SSD, the head is simulated and instantaneous. Switching the IO algorithm under Linux to No-op (first-come first-serve with no re-ordering) should allow the SSD to shine.

I found that with large SAN arrays and RAID array's with battery backed-up write cache, disableing IO scheduling optimizers increased performance significantly on database TPC tests and conversely killed performance on JBOD arrays with no cache. It stands to reason that the same effect would hold true on SSD.

I can't find any way to disable these types of IO optimization in Windows like in the example above.

Not 256 next year, but 64 now!

[rew]

I don't want 256Gb next year. I want 64Gb right now for about a quarter of the price.... I can manage with 40Gb on my laptop. 2Gb of MP3's is enough, 2-5Gb of operating system, a few movies, which have to be "replenished" at home after viewing. All in all, you can make a laptop work just fine with 32 or 64Gb. The biggest laptop drives are now 160 or 200Gb. So why go larger? Just provide bigger sequential speed, better random access speed, bump-proofness. Bigger size next year is fine.

HyperDrive Hybrid?

[Anonymous Coward]

When is someone gonna get off their duff and implement a hybrid RAM/flash drive with incorporated capacitor, where the RAM/flash ratio is very high, maybe approaching 1:1?

I mean, the HyperDrive4 is a desktop device, but fitted with a flash hard disk it turns into a monster. Condense the components to the density that cellphones are at now, and you probably have the god drive on your hands.

Smart controllers

[Joce640k]

If you fix things so that every 2Kb block on the disk can be remapped then those 2Kb writes can mostly be tagged onto the end of recently erased blocks, no erase needed.

(In practice NTFS usually uses 4Kb blocks so you'd optimize for that but the argument stands...)

This would also help a lto with wear levelling, etc. as you'd write the entire disk in a round-robin fashion, remapping blocks as you go.

The controller would need static RAM to hold the remapping table but that's no big deal these days.

I'm still waiting for real speed

[Heddahenrik]

This might be good news for laptops, but what I want is a PCIe-card with real speed compared to SATA-II.

I still haven't seen anywhere to buy it or heard anyone getting their hands on an ioDrive from FusionIO, but they are complaining about too high demand in Q2 (they started shipping it to big screaming customers the 7th of April).

But they have done it right: Straight into the PCIe and it comes with a Linux-driver. I just want to be sure about that it will do wonders for my databases before buying one. It's expensive, but compared to getting a machine with 60 GByte RAM or a SAN it's a bargain. And it's faster than any RAID-card, it's said, and you can RAID them if that isn't fast enough for you.

Isn't it possible to make one of these...

[Joce640k]

I'd like a USB dongle with, say, eight slots for SD cards in it. I can fill the slots up with cheapo cards and the controller would treat them like a RAID.

I'm sure performance would be competitive with a hard disk and it would only cost me $100 for cards to make a drive big enough for system, some workspace and swap file. Seagate's Raptor drives have had similar capacities and it hasn't held them back.

Re:Isn't it possible to make one of these...

[billcopc]

I started rolling my eyes after you said USB.

USB is useful, but it is not performant. I'll take an eSATA over USB, especially if RAID is involved.

The WD Raptor drives are falling behind, for several reasons. One is that "regular" drives are catching up to the Raptor's performance. The other is that often times, the limiting factor for performance is the motherboard's chipset. This has led to the true speed freaks (*blushes*) opting for PCIe SAS controllers with dedicated RAID processors. SAS drives can often be cheaper than Raptors, and yet offer better performance under seek-heavy loads which are common with today's horribly-coded games.

It's a game of numbers.

[Joce640k]

Have you seen the average notebook buy in a shop? It's all a game of "find the biggest number".

As a geek I'm always being asked if such-and-such a laptop is "fast enough", if XX is enough disk space, etc.

People have no idea what the numbers mean, or how they compare to the numbers six months ago. They don't even know the difference between RAM and hard disk. All they know is they don't want low numbers.

Bottom line ... "disk capacity" is a number on the little label so it has to keep increasing no matter what.

Re:Technology: Still new! Still Improving! Surpris

[Anonymous Coward]

"Solid State Drives for computers? They aren't really out of beta!" - by PC and Sony Fanboy (1248258) on Monday May 26, @09:18AM (#23543739) Ever heard of the CENATEK RocketDrive (PC-133 SDRAM 4gb onboard, PCI 2.2 interface)? Or, the GigaByte IRAM (DDR-RAM onboard, SATA interface)?? They've been around for YEARS now, & far from beta + they work (I have the former one since 2002 & works like a dream to this day - and, they're better than this FLASH-RAM based crap, because they use std. RAM that doesn't have write wear that burns them out, as do FLASH based "solid-state drives" (what a crock of crap it is calling them that, piggybacking on the good name of products that are TRUE SSD's, such as the ones I mention).