Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
IBM

Super fast storage access from IBM 57

Posted by Hemos from the i-want-it-right-now dept.
Several people wrote with the news burst from IBM annoucing a "storage breakthrough". Details are still forthcoming, but the advance will says it will be possible for computers to "store data 1,000 faster then they can now", however the actual usage of the product is some years out.
This discussion has been archived. No new comments can be posted.

Super fast storage access from IBM More

Super fast storage access from IBM

Comments Filter:
  • duh. Just go to the library and browse their copy for free. I haven't seen a library yet that doesn't carry Science.

  • Is it that far behind in R&D? I'd think not, since they have it, and it works, etc.. but, if they're claiming that much time to make it commercially available (how long does it really take to mass produce them?).. I wonder....

    Remember that the announcment of the technology and the announcement of the application of the technology are not the same thing. For all we know based on the released material the working technology demonstrator might be the size of a buick to hold all the supporting gear and require a supply of 10Kv @ 400GHz or something. Now they know they can do it, the challenge is to figure out how to put it on your desktop.

    Having said that, and IBM being who we are, I some how doubt if the researchers were allowed to even submit their paper to Science without having a very good start on turning this into a deliverable technology. (just imagine what they get to put in the "W" accomplishment section of their PBCs now.... :)

  • I'm getting tired of this hype.

    1,000,000 times more capacity!, 1,000,000 faster!,more robost than anything else!,.

    Oh...and it wont be availble for a couple of years. (i.e 7-9)

  • death to micro$~1!!!!!!
    "Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
  • Perhaps he intended the Randian definition of a Communist: Both taking and giving according to desires proclaimed as "needed for the common good". An accurate description of some scientists, who do not recognize the need for justifying the benefit of their research to garner high levels of public funding, and disdain private research because the source of the funding benefits.

    Of course, ScienceMag is run by the American Association for the Advancement of Science [aaas.org]. It's a non-profit group.
    What annoys me is that you have to join the AAAS and subscribe to the dead tree edition - which means both killing trees and supporting their lobbying efforts, some of which I object to.

    There's nothing wrong with paying money for a service (most of us would starve otherwise), but the reality is worse than the original poster felt. You have to hand over your ideology to gain access to the information - which defeats the purpose of getting information in the first place.
  • They're inexpensive compared to the 10 meg hard disk I got for my IBM XT many years back..
  • Do you compile all of your code with, at least, gcc -pedantic -ansi -Wall?
    Actually, I do. It makes porting stuff easier and you never know when that quick hack will become the core of a larger system.

    $.02
  • Actually, I saw this first on Anchordesk and came here becuase I figured I'd find more info than the little "news burst".

    I don't mind minor flatulence from ZD, but I would have expected the whole BM from /. (like a ling to Science article?)

    Oh well, it's years away anyway.

    Rick
  • the zd article was written before the information was released in sciencemag. but I still think I found the right stuff.
  • If I read one more gee-whiz story about some groundbreaking computer technology that's going to revolutionize everything in 3-5 years, I'm going to unplug my PC, move into the woods, and live on nothing but berries and 5.25" disks. If even half these announcements turned out to be accurate, we'd all be driving fusion-powered personal blimps and computing with superconductors and data crystals...
    Chances are, by the time IBM is actually bringing any of this stuff to market, some yet-to-be-invented startup company will already have rendered it obsolete.
  • "will says will says it will" ... "store data 1000 faster" ...

    1000 what faster? if "times", why didn't you just say so? Or were you waiting for Will to say it?

    --Corey (j/k)
  • Yeah, that would be smart of IBM. Except for the fact that IBM has it's own HDs, and probably would want consumers to keep buying them. This is a big breakthough, and IBM had every right to publicize it.
  • Um, you may not realize it, but Science is a real paper magazine also, and it costs money to print magazines. Even if it were only an online service, the people who maintain it have to make money somehow. If you don't pay people to keep the site up, what you will get is a crappy online magazine updated by random people in their spare time. I'd much rather have a high quality site that I have to pay for. In fact, I'd much rather just keep getting the paper version (which is much easier to read and doesn't crash) and ignore the web site.
  • Apparantly you have the impression that publishing
    material, on the web or in print, is free.

    That would be false.

    -WW

    --
    Why are there so many Unix-using Star Trek fans?
    When was the last time Picard said, "Computer, bring
  • This article [sjmercury.com] on SiliconValley gives a tad more information, though it's still lacking any real content.
  • This would then be precisely the same effect as GMR in reverse. A spin dependent tunneling junction could act as a memory. How is the current being applied then. My guess would be through an AFM/STM. IBm spent a _lot_ of time and money on spin polarised tunneling, and then it diappeared. A single read write head with an integrated SPSTM and MFM or GMR read head would work. This would integrate well with their work on large scal AFM integration.

    Sorry about all of the acronyms. Too use to preaching to the choir. AFM is Atomic Force Microscopy--see http://www.di.com/products/ScanMethods/ScanMain.ht ml
    STM is scanning tunneling microscopy. see
    http://www.almaden.ibm.com/vis/stm/ if you want to see some marvelous examples. IBM did invent the STM.

    Finally, if this supposition is correct, then this isn't anything new. MRAMs are based on exactly this principle. I think it has been demonstrated with CMR (colossal magneto-resistive) materials, as well as with spin valves. An INSPEC search finds many of them
  • "If even half these announcements turned out to be accurate"

    That's how I felt about the announcemetn that they had technology that would make multi-gigabyte drives smaller and cheaper; and sure enough, just
    a few years later, that is the case (and I'm still
    tempted to marvel at a 10 gig laptop drive)
    Those things qualify as inexpensive, consumer-market items, right?
  • minor nit:

    If the scientists are communist, that would mean they would be taking according to their needs (building upon the work of others) and giving according to their abilities (publishing the results).

    The scientific comminity is, perhaps, no longer "communist", but rather is transforming itself into an elite club of money grubbing capitalist bourgeois.

    :-) :-) :-)
  • Actually, this [sciencemag.org] is what ZDNet was talking about. Note that three authors are from IBM: Science Aug 6 1999: 864-867

    Minimum Field Strength in Precessional Magnetization Reversal

    C. H. Back, 1* R. Allenspach, 2 W. Weber, 1 S. S. P. Parkin, 3 D. Weller, 3 E. L. Garwin, 4 H. C. Siegmann 1

    1 Laboratorium für Festkörperphysik, ETH Zürich, CH-8093 Zürich, Switzerland.
    2 IBM Research Division, Zurich Research Laboratory, CH-8803 Rüschlikon, Switzerland.
    3 IBM Research Division, Almaden Research Center, 650 Harry Road, San Jose, CA 95120, USA.
    4 Stanford Linear Accelerator Center, Stanford University, Stanford, CA 94305, USA.

    Ultrafast magnetic field pulses as short as 2 picoseconds are able to reverse the magnetization in thin, in-plane, magnetized cobalt films. The field pulses are applied in the plane of the film, and their direction encompasses all angles with the magnetization. At a right angle to the magnetization, maximum torque is exerted on the spins. In this geometry, a precessional magnetization reversal can be triggered by fields as small as 184 kiloamperes per meter. Applications in future ultrafast magnetic recording schemes can be foreseen.
  • You're all fucking idiots
  • Where did that article about deep-linking movie-trailers that was posted a couple of hours ago go?
  • so where are the specs??? I wanna see one.

    Show me show me!!!

    Cynic??? Who's a cynic??
  • If this is the correct item, then it is what would have made either the ZD posting or the /. posting into something worthy of web space.

    Fluff is pervasive, and in the end, is directly analogous to spam.
  • More information would have been more appreciated, and I agree that the link provided really doesn't say a hell of a lot, but the link WAS referred to as a "news burst". That can hardly be classified as an "article". It wasn't mis-represented. However, IMHO, it shouldn't have been posted unless there was some real meat to it.

  • I just got some
    very weird looks from colleagues wondering what was so fucking funny... Someone moderate that up. :-)

    D.
    ..is for Doh!

  • from sciencemag, today (Score:4)

    by lcracker ( 10398 ) on Friday August 06, 1999 @06:32AM (#1761336) Homepage
    [i think this is the what zdnet was talking about]

    Magnetic Switching with Current Pulses

    The switching of magnetic memories usually requires application of external magnetic fields. Two reports show how electric current pulses within small devices can be used to
    switch magnetic states. The rate at which magnetic data can be stored is limited by the time taken to reverse the magnetization orientation. In conventional methods, the reversal
    field is applied antiparallel to the out-of-plane magnetization, and the switching speeds are limited to the nanoseconds. Back et al. (p. 864) show that this limitation may be
    overcome by switching the magnetization of thin, in-plane orientated ferromagnetic films using ultrafast focused current pulses from an accelerator. Although the magnetic fields
    generated were no greater than those of conventional recording heads, reversal times faster than 5 picoseconds were demonstrated. The electrical resistance of stacked layers of
    magnetic (ferromagnetic) and nonmagnetic materials depends greatly on the orientation of the magnetic moments in adjacent magnetic layers--if the moments are aligned, the
    resistance is low. Myers et al. (p. 867) show that the magnetic orientation in a single layer can be switched to be aligned or misaligned to a second magnetic layer by controlling the
    direction of an applied spin-polarized current pulse. Such switching may find application in fast, nonvolatile magnetic memories.

  • A better url (Score:1)

    by Anonymous Coward
    This [sjmercury.com] site has a little more information/details.

  • What sick construction is this?

    Do you compile all of your code with, at least, gcc -pedantic -ansi -Wall?

    I mean, ALL of it? Just asking....

  • wtf? (Score:1)

    by crayz ( 1056 )
    What are you talking about? How is it not yet news? Just because they aren't shipping them doesn't mean it isn't news. I guess if scientists said that a comet was going to destroy Earth in two months, that wouldn't be news either. It'd only be news after the comet had hit.

    Access speed for storage is very important, and is quickly becoming the bottleneck w/ new computers. If IBM found a way to totally eliminate that bottleneck, I consider that to be big new indeed.
  • C'mon, guys. "News for Nerds. Stuff that matters." This thing is not yet news, and is not real enough to be "stuff" at all, much less "stuff that matters".

  • Actually, I do.

    :) I figured *someone* would call me on that... I would bet that you don't have perfect grammer either... Nor someone to proofread every post to slashdot...

  • You're all fucking idiots

    As the article in today's Slashdot demonstrates, we may be idiots, but we sure as hell ain't fucking.

    Kaa
  • Ok, you got me there. But my answer to your question still stands.
  • Touche! :)
  • It's news, not history, dude. Really. News is bleeding edge. Don't like it, move on . . .

    later
  • So I browsed out to Science magazine, and after the third degree to sign up for a "limited membership" they dont really give you access to useful info.

    What is science coming to when we have to pay to read what new discoveries are out there? what kind of sick and twisted corporation puts a price on knowledge??? So much for freedom of information....

    The scientific community is turning into a bunch of money grubbing communists...

    Cynic??? Who's a cynic?
  • This is news.

    The thing is, I wonder why it will be so long before it's commercially available, and yet they're posting news briefs about it, they have a paper out about it, etc.

    Is it that far behind in R&D? I'd think not, since they have it, and it works, etc.. but, if they're claiming that much time to make it commercially available (how long does it really take to mass produce them?).. I wonder....
  • What ever happened to putting more read/write
    head/armatures on hard drives?
    I mean.. duh. two head opposite sides of the platter- one writes the other verifys, - single
    platter spin.

    I remeber seeing one hudge drive opened up
    at DiskDrive Depo on day.
    Even if the drive sizes go back up to 5.25
    form factor, I'd be happy (member bigfoots?).
  • Science magazine is a pay subscription site. Boo! I guess probably no cypherpunks/cypherpunks for us.

    You can peruse the table of contents for free though, and the article is currently on the site.

Slashdot Top Deals

He has not acquired a fortune; the fortune has acquired him. -- Bion

Close