Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
Data Storage

IBM's Optical Storage Is 50 Times Faster Than Flash, And Also Cheaper (prnewswire.com) 77

Flash storage is not as fast as the main memory (RAM); but RAM can't be used to store your regular files because of its volatile nature (and also because it's expensive). It appears we may soon have the perfect middle ground of the two. Scientists at IBM have demonstrated reliably storing 3 bits of data per cell using a relatively new memory technology known as phase-change memory (PCM). Engadget reports: To store PCM data on a Blu-ray disk, you apply a high current to amorphous (non-crystalline) glass materials, transforming them into a more conductive crystal form. To read it back, you apply a lower voltage to measure conductivity -- when it's high, the state is "1," and when it's low, it's "0." By heating up the materials, more states can be stored, but the problem is that the crystals can "drift" depending on the ambient temperature. IBM's team figured out how to track and encode those variations, allowing them to reliably read 3-bits of data per cell long after it was written. That suddenly makes PCM a lot more interesting -- its speed is currently much better than flash, but the costs are as high as RAM thanks to the low density.
This discussion has been archived. No new comments can be posted.

IBM's Optical Storage Is 50 Times Faster Than Flash, And Also Cheaper

Comments Filter:
  • So it's cheaper (Score:5, Insightful)

    by goombah99 ( 560566 ) on Tuesday May 17, 2016 @03:13PM (#52129453)

    in the sense that it costs more?

    • That *is* odd... but perhaps it's because they just can't even.
      • by duckintheface ( 710137 ) on Tuesday May 17, 2016 @03:31PM (#52129615)

        Despite the use of the term "optical" there is nothing optical or "blue-ray disk" about this. Calcogenide glass is the PCM material. It is written and read with the application of voltage. There is no spinning disk involved.

        The most obvious omission is a comparison to the Micron/Intel 3D Cross Point memory announced last summer and scheduled for commercial introduction in 2017. 3DXP is 1000 times faster than flash (not just 50 times faster). There would also seem to be a number of patent issues since 3DXP also uses calcogenide crystals as the storage medium.

        It is journalistic malpractice to write an article like the two linked here without comparing the IBM research to the previously announced work by Micron and Intel.

        • Can I sue for "journalistic malpractice"?

        • "It is journalistic malpractice..."

          Well, first we would have to have unbiased journalists who have wide and deep knowledge of their subject matters, are not beholden to "editorial content standards" and not ideologically or financially bound to...I think I see the problem.
    • Re:So it's cheaper (Score:4, Interesting)

      by rahvin112 ( 446269 ) on Tuesday May 17, 2016 @03:27PM (#52129569)

      About the same as claiming it's an optical format when it's actually phase change. Throw those buzz words in.

      Phase change memory already exists as Intel/Micron XPoint memory which should go on sale later this year (it already been sampling for a few months).

    • in the sense that it costs more?

      They make up for it in volume.

      Math!

  • Whoa. (Score:5, Funny)

    by charlesbakerharris ( 623282 ) on Tuesday May 17, 2016 @03:14PM (#52129461)
    That must make it, like, 500 times faster than Silverlight.
  • by Anonymous Coward

    "amorphous (non-crystalline) glass"

    So a glass, glass-glass then?

  • by Art Challenor ( 2621733 ) on Tuesday May 17, 2016 @03:25PM (#52129551)
    Finally! Ternary Digits, now I won't be twiddling bits all day.
  • This has been researched since the 1960s, but it just was never viable for commercial production. Yes, this is a mojor breakthrough, but it's facing heavy competition. Intel/Micron's new memory tech will be out and it has already been demonstrated to be 1000 times faster than flash.
  • by Anonymous Coward

    The 3d xpoint resistive memory being developed has higher density than flash, should be manufacturable with modest changes to exising fab methods, has no write life limitations, is way faster than flash, does not need to be updated in large blocks like flash, does not use exotic materials, and should be cheaper than flash once mature.

    Why are we considering physical phase change materials that have potential write life issues from crystal grain migration/unexpected crystal growth over the cell domans, when w

  • Website about the research(note, there is a list of papers too... this "news" is actually from 2015) https://www.research.ibm.com/l... [ibm.com] Youtube video: https://youtu.be/q3dIw3uAyE8 [youtu.be] And yes... they only have a prototype. There are still lots of possibly, maybe, somedays in this announcment.
  • PCM is not limited to re-writable Blu-Ray, it is actually used in memory chips.

    Micron used to manufacture PCM memory chips and dropped them in 2014 [globalspec.com]. There are also some debate regarding whether IMFT's 3D XPoint [micron.com] is also PCM or not [theregister.co.uk].

    The real innovation in IBM's work is turning PCM into a TLC, and that is really impressive.

    • PCM in blu ray discs stands for "Pulse Coded Modulation" while PCM in memory chips stands for "Phase-Change Memory". These are not the same PCM, whoever wrote the summary is very very confused.
  • I am not sure what tier this storage goes in, as for price/capacity, where CPU registers are the most precious and tape, cloud, or big, slow DASDs are on the other scale of speed/price.

    Does it go under flash, but "above" spinning rust? The BD example made it seem that IBM is wanting to make another CD-PD, which was a storage format that lasted a few years until CD-R and CD-RW became mainstream, where a drive could read CDs, as well as use a specific optical cartridge.

    Personally, I would love a high densit

    • by swb ( 14022 )

      I get the impression that this and IntelMicron's 3DXpoint are really meant to be replacements for current uses of SSD and most uses of disk.

      I have the vague feeling that maximal price extraction will see 3DXpoint used in some cases as a caching layer between CPU and flash disk initially until manufacturing capacity ramps up and the production cost gets low enough to drive SSD flash out of the market, which it ought to based on performance alone.

      Hopefully it will drop in price enough to kill off hard disk co

      • PCM won't kill flash (3D NAND has some cost benefits), it will displace DRAM in some applications.

        And I would not count HD out as well, there has been some impressive advances published lately by SanDisk and the like.

        Computerworld published recently a nice article on the Memory landscape [computerworld.com]

        • by swb ( 14022 )

          I think PCM will have to cut costs to become price competitive with 3D NAND as primary storage in order to survive. I don't think it's cost or value proposition as some kind of buffer technology or DRAM augmentation will get it very far.

          For better or for worse, the entire technology field is stuck in a DRAM/disk paradigm and no emerging storage technology (as in likely to be mass produced for consumption in 5 years) will be able to unify CPU and disk storage. If 3DXpoint lives to 100% of its hype, it migh

      • by mlts ( 1038732 )

        I have been asking the same thing. The LTO drives have already dealt with shoe-shining, so they can work on a sub-optimal interface, and even then, there is always having the drive with an I/O buffer similar to CDs before "burn-proofing" made buffer underruns a thing of the past.

        I do know LTO-3 drives can run from Thunderbolt 1, which is about on par with USB 3, so it can be done. It is just a matter of making consumer level software to handle LTFS, and some basic documentation, like not to run a defragme

        • by swb ( 14022 )

          The thing is USB3 and 3.1 really aren't sub-optimal interfaces. USB 2 was, but 3 eliminated the interrupt problem and took it to 5 Gbps and 3.1 to 10 Gbps.

          With this kind of bandwidth, I would kind of hope manufacturers would look to expand market penetration on some "enterprise" devices and realize that between the inexpensive, high speed connectors and operating system device virtualization that more devices could be effectively connected via USB3 without SCSI as a dependency.

  • It sounds cool, but how does it incorporate lasers and virtual reality while traveling at warp speed?
  • One of the few techs to actually hit market within about 10 years of being proven feasible.

    It's good stuff, but this stuff is slow in comparison to Intel/Micron's stuff.

    If IBM partnered with AMD on making a 3D stacked version of this, they could easily catch up or even be faster, with higher bandwidth to boot.

  • by friesofdoom ( 3817155 ) on Tuesday May 17, 2016 @04:40PM (#52130137)
    "To store PCM data on a Blu-ray disk, you apply a high current to amorphous (non-crystalline) glass materials, transforming them into a more conductive crystal form. To read it back, you apply a lower voltage to measure conductivity -- when it's high, the state is "1," and when it's low, it's "0.""

    1. We do not write to blue ray disc by applying a voltage, we shine lasers at it.
    2. The summary seems to have confused the fact that Chalcogenide is used in both RW Optical media, and phase-change memory.
    3. The summary thinks that IBM has invented a new optical memory, when they are clearly talking about phase-change memory.
    4. The summary has confused two acronyms of PCM - Phase Change Memory and Pulse Coded Modulation.

    This whole summary is utterly cringe-inducing and complete garbage. Sorry.
  • Like to have non battry backed ram temp disk that does not need to even have all matched speeds.

    That can make good use of all older ram that people have laying around.

  • HP's memresistor project "The machine" is going to so thrash this technology to hell and back. They are talking about petabytes of data as fast as ram. You won't need to have a separation between ram and disk anymore. They already have prototype and they should show up in a few years.

    http://fortune.com/2016/03/02/... [fortune.com]

A committee takes root and grows, it flowers, wilts and dies, scattering the seed from which other committees will bloom. -- Parkinson

Working...