IBM will kill the hard drive in ten years
Seems only fair seeing as it invented them
IBM BOFFINS reckon that a new kind of storage device with the speed and efficiency of flash, and the capacity and low cost of platter drives, could be less than ten years way.
Both prevalent technologies have their problems. HDD is hot, noisy and bulky. It's also sensitive to knock and bumps, and has loads of moving parts to go wrong. SSD is smaller and lighter with no moving parts but, despite being quick at reading data, is a bit slovenly when it comes to writing it. SSDs also have lifespan problems, as each write cycle does tiny amounts of damage to the integrity of the circuits. Oh, and it costs one arm + one leg.
Dubbed "Racetrack", because the data races around a wire track, IBM's new baby has no moving parts and potential capacity, size for size, more than 100 times greater than current technologies.
It would also require far less power and generate far less heat meaning that battery-operated devices could run for weeks on a single charge and last for decades.
IBM is no noob when it comes to bringing revolutionary concepts to market having invented the memory chip and the hard drive.
IBM fellow Stuart Parkin and a number of colleagues published their findings in this month's Science. In their paper, the scientists describe their use of horizontal permalloy nanowires 10,000 times thinner than a human hair charged with spin-polarized current pulses to record and retrieve data.
We reckon describing the new chips as "virtually indestructible" might backfire on them, though. Remember when they said that about CDs?
L'Inq
IBM
Comments
IBM memory
I read that article, very interesting. But when they discuss the "destructability" of the memory, they are talking in terms of a comparison to today's devices. They don't mean literally.I can't link directly to the article, as you need to be a member, but I can link to the registration page that let's one register for free to see the weekly updates. The name of the article is "Magnetic Domain -Wall Memory.
http://www.sciencemag.org/content/vol320/issue5873/twis.dtl
'Bout Time
We reckon describing the new chips as "virtually indestructible" might backfire on them, though. Remember when they said that about CDs?Still its a step in the right direction.
The platter based HDD drive hasn't really changed at all since IBM first produced them. Ok they've come down in size and have increased in capacity, but the performance of HDD's have not improved relative to the rest of the hardware in a modern day computer.
It truely is about time a new tech was discovered and rolled out and yes Flash is again a step in the right direction but the price is astronomical, and as you say the lifespan is limited.
I welcome the news of the research and look forward to some real world benchmarks.
Science = Weekly
Science is a weekly journal so the phrase " this month's Science" is a bit off.The article itself is not very much of a scientific paper. It is a blurb about how the project seemed to nave controversial funding and focused a lot on a naysayer. As with all nanotech it will take a real major breakthrough to get the fabrication to work on the scale necessary to go into mass production. I wish them luck because it is an interesting technology.
Its deja vu all over again
Memory that consists of tiny magnetic domains circulating on a solid-state chip?Almost like... little bubbles.
Hey, they could call it "bubble memory"!
I wonder if I should trademark that name before anyone else thinks of it?
What goes around comes around
What is absolutely amazing about this is its similarity to the mercury and wire delay lines used in the memories of the very first computers. History repeats itself:http://en.wikipedia.org/wiki/Delay_line_memory
Hype-R-Drive
This sounds just like another technology I've heard of - it's called the hyper drive. it's all hype until i can purchase the drive.Delay Line Memory?
So is this a refresh of delay line memory, basically?Why wait 10 years?
The PCIe x4 FusionIO drive is out late May 2008 in retail shops (but can be purchased now)It does not wear out like SSD, and is alot faster!
less limited than the i-Drive from Gigabyte.
http://www.youtube.com/watch?v=o-8iXADhfCE
I hope this replaces HDD.
sources????
"SSDs also have lifespan problems, as each write cycle does tiny amounts of damage to the integrity of the circuits."This is old-school thinking.
Is this the BBC now or something?
Better L'inq, please
Nice to see references in your articles, but please cross-check your references, too. The author of the cited article is too dumb to spell the name of Nobel Price winner Peter Grünberg correctly. Those kinds of articles may be good enough for the audience of CNN, but not the readers of The Inquirer.Uh Oh...
Uh Oh, Seagate may sue IBM when they'll launch it :Phttp://yro.slashdot.org/article.pl?sid=08/03/24/141212&from=rss
HDD Performance
Chris said:"[...] the performance of HDD's have not improved relative to the rest of the hardware in a modern day computer".
Really? Have you done the math before saying that? Well, let's see, shall we?
Benchmarks I have here say that a modern 320GB Seagate (ST3320620AS) has 64134.5KiB/s average read speed. A newer 1TB drive will be faster, but I don't have those numbers, sadly.
While I understandably don't have such exact data, I remember my first HDD way back at the dawn of civilization in A.D. 1996 (a 245MB Maxtor) was capable of something like 300KiB/s.
That's about 218-fold (or about (2^7.740)-fold) increase in performance.
Moore's law, if you recall, calls for a doubling in performance every 18 or so months. So we multiply and get 11.61 years. Add to 1996 and you get 2008. What a strange coincidence!
;-)
FYI, HDD and memory capacity are ahead of the curve. Processing power is hard to quantify: how many 486DX2s is a single Q9450 worth, processing-wise, anyway? But I have no reason to expect it to be falling behind.
And they say Kurzweil is crazy...