Disk Drives Face Challenge From Chips

WSJdpatton writes "Researchers are reporting significant progress in perfecting a different way to store data in semiconductors, which could replace one widely used type of memory chip and possibly become a credible competitor to disk drives. The researchers, in a paper being delivered at a technical conference in San Francisco, say they used a novel combination of materials to create prototype phase-change components that are more than 500 times as fast as flash chips, while requiring less than half of the electrical power to record data."

Re:Good news

[tttonyyy]

Arguably, this is the important part, and one reason why Flash would never have been a good replacement for a HD even if the speed issues were resolved:

Flash memory is popular because it retains data without a constant electric charge. Such chips aren't usually used in place of disk drives, because of their higher cost and because there are limits on how many times data can be written. Phase-change memory doesn't have that problem

(emphasis mine)

Re:No more harddrives?

[Daniel Dvorkin]

I don't want to think about the cost of 4G of this stuff though. *shiver*

[shrug] A decade ago, I'd never even seen a machine with 4GB RAM, and five years ago, I'd only ever seen that much RAM in monstrously expensive servers. Now I have a machine with that much RAM on my desk. (And yes, I use it; most of my work is pretty heavy number-crunching.) So if this stuff turns out to be viable, it'll get there.

Actually, a better comparison just occurred to me: about fifteen years ago, I paid an extra thousand bucks to get a laptop with a 60MB hard drive (vs. the standard 20MB or whatever it was). A few months ago, I bought a 256MB thumb drive for about twenty-five bucks. That just blew me away when I thought about it.

Re:No more harddrives?

[Ngarrang]

I don't think the hard drive will disappear completely, but as the costs come down, the companies cannot make money producing the smaller capacity drives. We will see 1Tb hard drives readily available someday, sure thing. But different people have different needs. Hard drives are beginning to augment backup strategies because they have become so cheap and high in capacity.

A solid state drive has a higher G-shock tolerance, is quieter and requires less power than a hard drive. These features are why the technology is attractive to the people who need it. And not everyone needs a hard drive that is 400gb in size. Network appliances may only need a small 1gb boot drive, and these kind of devices will need this new phase-change memory, or whatever will work for the task beyond flash.

It would be cool to have something like this that is your main memory AND your storage space in one. We could call it Run-In-Place. We could then have a instant-on computers. Just imagine Windows XP or Linux booting up in under 3 seconds!

Re:The real challage is price.

[stevesliva]

If price wasn't a case Computers wouldn't have much RAM but all Cache, or huge amount of registers. But in real life price is the final decision

Actually in systems where price is no object, performance is usually paramount. If you have astounding amounts of registers or cache, your performance per instruction or memory operation may be slower. Given the fact that we can manufacture dual-core dies with ease, I imagine we could easily fit a bazillion more registers or double the L1 cache of a single core, but there is a performance trade-off there.

Re:Good news

[el_womble]

Do you really believe that harddisks don't fail?

The difference is that flash fails with writes (not reads) and HDD fails with reads AND writes (bad sectors?). Early flash failed after only 10,000 writes per sector, newer flash is in the millions. Flash spreads the writes around, so to reduce the chance of any one sector failing and can do this because flash is genuinely RAM (unlike HDD where location affects transfer speed). Both HDD and SSD employ firmware stratergies that hide sector failure from the OS, only flash can do that without any real cost to performance.

The end result is that if either are working after 3 or 4 years your doing well, and should really be looking for a replacement unit.

Re:Good news

[morgan_greywolf]

Do you really believe that harddisks don't fail?

No, but HDDs are amongst the most reliable storage media. A good, well-built SCSI drive can last for much, much longer than 3-4 years. I've personally seen hard drives as old 10 years functioning without a hitch. RAID can very much mitigate the risks associated with keeping drives around that long, too.

Moore's Law and disks

[billstewart]

For the most part, disk capacities have been increasing faster than the Moore's Law double-in-18-months for the last few years. I stopped caring about disk capacity somewhere around the time 6GB drives got replaced by 20GB drives which got replaced by 120GB drives over about 2-3 years, each at under $100/drive. (Then I got BitTorrent and started downloading lossless-compression music, so I temporarily had to pay attention again :-)

My first Vax, 22 years ago, had 1GB of disk, in the form of four washing-machine-sized drives which used removable 250MB disk packs. The drives cost about $120K total, and the packs were about $1000 each. There isn't really an exact comparison to that combination; you could either look at DVD-RW ($40 for the drive, $0.50 for the disks, so 8-12000x the price/capacity), or amortize the drive across some number of packs to compare to fixed disks (e.g. 10 packs per drive would be $160K for 10GB, though I think we only bought about 3 packs per drive over before that machine was obsolete), or you could make some unbalanced comparison like $20 for a CF-to-USB adapter and $20/GB for Compact Flash cards, which would be a mere 200:1 on the removable media but 6000:1 for the "drive".

Re:Moore's Law and disks

[DarkSarin]

And I have an astrophysics friend who just told me that the university we attend (grad students) will be putting up a new satellite that can generate 30TBs of data in one night (Full Sky Scan!). Wow.