Forgot your password?
typodupeerror

IBM and Fuji Announce Tape Storage Breakthrough 254

Posted by ScuttleMonkey
from the data-in-a-volkswagon dept.
robkill writes "IBM and Fuji have announced a breakthrough in the amount of data that can be stored on magnetic tape, a 15X improvement to 6.67 billion bits of data per square inch. IBM estimates that it will be 5 years before this hits the mass market"
This discussion has been archived. No new comments can be posted.

IBM and Fuji Announce Tape Storage Breakthrough

Comments Filter:
  • Death? (Score:2, Funny)

    by UnixSphere (820423)
    When will tape die?
    • Ask Netcraft...
    • Re:Death? (Score:3, Funny)

      by Draconix (653959)
      When someone walks by with a magnet in their pocket?
    • Re:Death? (Score:5, Insightful)

      by Ryan Amos (16972) on Wednesday May 17, 2006 @06:48PM (#15354753)
      When a 500 GB hard drive costs $75, can be thrown across the room and have a chance of working, weighs the same as a tape and can be easily inserted/removed in bulk with software management and barcode readers to keep track of it all for you.

      Until then, tape will stick around. I have a feeling it might be a while.
      • We're honestly not that far away from $75 500GB hard drives. Another 2-3 years, perhaps.

        I'm honestly surprised that the state of optical media has progressed so slowly though. BlueRay and HD may seem very large, but considering the size of our hard drives, I'd be happier if 5 inch CD formfactor media could store on the order of ~100GB.
        • Re:Death? (Score:4, Insightful)

          by TeknoHog (164938) on Wednesday May 17, 2006 @07:13PM (#15354870) Homepage Journal
          I'm honestly surprised that the state of optical media has progressed so slowly though. BlueRay and HD may seem very large, but considering the size of our hard drives, I'd be happier if 5 inch CD formfactor media could store on the order of ~100GB.

          There's also some advantage in separating the storage medium from the read/write heads. If either part in a hard drive fails, you're literally fscked (except for some really expensive recovery solutions by Ibas [ibas.com] or the like). On the other hand, you can always put an optical disc in a brand new drive. And if a disc is scratched beyond readability in your current drive, chances are you can read it with another drive in the future.

        • "Another 2-3 years, perhaps."

          Probably not even that. The increasing simplicity of disk mergers and unconventionally attached storage will drive the larger capacity disks into the region of best price per GB faster.

          "I'm honestly surprised that the state of optical media has progressed so slowly though."

          Personally, I'm not so much surprised as I am disappointed. It was more or less obvious it would go this way once you realized the companies were too deeply tied to the film industry. By the time they become p
      • When a 500 GB hard drive costs $75, can be thrown across the room and have a chance of working, weighs the same as a tape and can be easily inserted/removed in bulk with software management and barcode readers to keep track of it all for you. As far as the price is concerned, you're probably looking at a couple of years from now (and this tape isan't available now, either).

      • Re:Death? (Score:3, Insightful)

        by dgatwood (11270)
        The problem with tape is that the capacity sucks relative to what you're backing up. The biggest tapes (as best I could find in a quick web search) hold 400GB each. The hard drive in my desktop is 500GB. My desktop's drives total 750GB. It would take two full tapes to do a single full backup.

        The industry predicts that with the newer drive head technologies coming out, HD capacity will double every 12 months. This means that:

        1 year: 1 TB
        2 years: 2TB
        3 years: 4 TB
        4 years: 8 TB
        5 years: 16 TB

        So wi

        • 400 GB takes 83 square feet.

          93, if my math is right. Anyway, you slightly understate the problem. Multiplying a 680m 400GB tape by 15 means you'll get about 6TB on that same 680m tape, which is less than 50% of your projected 16TB hard drive - less than 40%, in fact.

        • Re:Death? (Score:3, Interesting)

          by zippthorne (748122)
          So.. you don't spool your tape? You'd just kind of have it lying around.. flat.. on the floor?

          Who cares about the AREA required. We don't live in flatland. Tape is freaky thin. What is the VOLUME required for all that storage.
        • Re:Death? (Score:3, Interesting)

          by bhiestand (157373)

          The biggest tapes (as best I could find in a quick web search) hold 400GB each.

          Let's just say your web search was incomplete. About four years ago I worked for a company that made parts for 1TB tape drives. I know they have at least 8TB tape cartridges developed now. I'm sure the price per byte is more expensive than modern, cheap consumer hard drives, though. Keep in mind you're comparing apples to oranges, and essentially complaining that it's much harder to peel an apple. Companies that need large

      • When a 500 GB hard drive costs $75, can be thrown across the room and have a chance of working

        I have yet to see a tape that could survive being thrown across the room.

      • I agree with most of what you said, but don't understand why the point of adding the bardcode bit at the end was for. I mean it's not like you can't put barcodes on hdds to be scanned.

        However, this introduces a different problem. 500GBs will one day reach that pricing, no doubt about that. But when it does, tape will be even cheaper and probably more dense.

        What will eventually replace tape is not hdds, but either flash memory or optical discs. Flash and optical disc both have a speed advantage (especially f
        • I agree with most of what you said, but don't understand why the point of adding the bardcode bit at the end was for.

          With barcodes, you can scale up your system with a tape storage robot [ibm.com]. The barcode is mostly there for error checking to make sure a tape is in the correct spot. When I worked at the PSC [psc.edu], they had two fairly large tape robots. I had to rewrite the barcode generator program, so that it no longer depended on a library with a strict beerware license [wikipedia.org] (No, I did not make this up).
          • but what I don't get is why is this something pertaining only to tapes. if i had a hdd robot, an optical disc robot, or a flash memory robot (i know only of the optical disc robot existing, not sure about the others, but i don't see what's stopping them from doing that if they really want to automate the barcoding), i can just as easily barcode them.

            the other disadvantages were technical disadvantages, while the barcoding is definitely a disadvantage that could definitely change. invest some money, make you
      • I guess you've never dropped a DLT tape 2 feet onto a carpeted floor and have it stop working like I have. And yes, I tried to retension it too.

        The point with harddrive storage is that you don't have to move them... ever. Run your backups over the Internet to your remote sites. Having everything on-line makes the management a lot easier too. No fooling with barcodes and robots either.

    • When will tape die? - immediately after you write to it.

      Tape is a write only medium. Upping densities will just make it worse.
      • When will tape die? - immediately after you write to it.

        Tape is a write only medium. Upping densities will just make it worse.


        Sounds like you have only used consumer- and SOHO-grade tape drives.

        Business-class drives and media like (S)DLT, LTO & older formats like 3490 are very durable.
        • I don't think he was commenting on the durability, but the linearity. Since you can't access arbitrary positions on the tape without spooling through all the intervening tape.
          • When will tape die? - immediately after you write to it.

            I don't think he was commenting on the durability, but the linearity.


            Really? If that was my complaint, I'd have written something like:
            When will tape die? - As soon as you want to retrieve the last file on the tape.

      • Only if your backup admin is criminally incompetent. Which unfortunately many are.
    • When all the governments of the world cease to exist. We maintain specific types of tape drives just because of different agencies we have to deal with. We have some "lowest denominator" type drives just in case. (lots of fun when one file can span a hundred tapes and the dorks that want it only can use that tape type)

      We keep end of quarter and end of year tapes till they no longer function, and then we keep them because we are too afraid to rid ourselves of them.

      Tapes are actually one of our lower cost
      • We keep end of quarter and end of year tapes till they no longer function, and then we keep them because we are too afraid to rid ourselves of them.

        Could you ditto the files from the old tapes to new tapes?
    • When will tape die?

      When will die tape?
  • That... (Score:5, Funny)

    by remembertomorrow (959064) on Wednesday May 17, 2006 @06:43PM (#15354723)
    is a lot of porn.



    What?
  • Funny how the article doesn't specifically mention actual storage capacity... Just vague physical dimensions.
  • by blair1q (305137)
    How long does it take to write this stuff?

    And how long to seek?

    Because if it isn't faster than swapping old-technology tapes, it's not worth a damn.
    • You'd usually use tapes for archiving, not stuff you'd access alot.
    • In fact without a fiber connection the drives we have would outrun and scsi connection we have available.

      Here is a good page with some information about different offerings from IBM from the last 20 odd years.

      http://www.dpts.co.uk/hdm02.htm [dpts.co.uk]
  • 5 years? (Score:2, Funny)

    by the_humeister (922869)
    By then my collection will have surpassed this capacity...
  • I read about it a history book a few years back (right next to zip disks and 8 track players)....

    What's next? Commodore 1024?
    • Tape is still a very reliable, relatively durable storage medium for taking your backups off site. At my office, we're currently debating the relative merits of tape vs. removable hard drives for our next upgrade. Removable drives are starting to look good, but it's by no means a one-sided contest. Tape is still used, and will be for quite some time.

      • The beauty of tape is you only need one drive, and can store huge (huge) libraries of data. I've seen one installation in a datacentre where the tape racks literally had a robotic arm picking the appropriate tape, moving it from the fireproof vault to a seperate drive, inserting it, doing the backup/restore, then replacing the tape. Thousands of TB of backup data there.

        Admittedly for smaller servers a removable HDD may prove far better given the usage vs. cost, but for large installations it's near impossib
  • leapfrogging (Score:5, Informative)

    by ziegast (168305) * on Wednesday May 17, 2006 @06:50PM (#15354762) Homepage
    FYI: Sony claimed [aittape.com] 11 billion bits per square inch quite some time ago.

    It's always good news when someone figures out how to store more bits into the same amount of space, and I'm sure that companies like IBM and Sony will keep pushing the limits.
  • by HardCase (14757) on Wednesday May 17, 2006 @07:07PM (#15354842)
    So if you combine the tape density breakthrough with the Linux device driver breakthrough, can you go faster than the speed of light? Or does SM just need a thesaurus?

    -h-
  • How many years? (Score:3, Insightful)

    by Nom du Keyboard (633989) on Wednesday May 17, 2006 @07:17PM (#15354891)
    Isn't five years the typical projected time span for something you're not going to see nearly as soon as you think? I remember in five years we're going to have flat-screen televisions you can hang on your wall. And while we do have exactly that now, I first heard this prediction twenty-five years ago!

    And I'm still waiting for the flat-screen TV you unroll like a poster and tack up with some double-sided tape.

    So this IBM announcement fails to excite. Five years is a very long time in the technology industry.

    • I'd just like to point out that 5 years predictions (and plans based on those predictions) worked stunningly well for quite a few communist countries.
  • by csoto (220540) on Wednesday May 17, 2006 @07:18PM (#15354899)
    All you naysayers, understand that we had exactly the same sort of announcements before the Linear Tape Open (LTO) standard was developed. IBM led a group of manufacturers to develop a standard built around a few breakthroughs in tape density and drive head technologies. They predicted 10X (or more) capacity, 5X (or more) throughput, etc. and it would be available in 5 years or so. Sure enough, LTO-1 came about and immediately led to a tape storage boom. Quantum pushed DLT to about its limits, Storagetek upped the ante with their very high speed formats, etc. Everything got cheaper. Tape stayed relevant. I predict the very same trend in the near future...
  • The MPAA is asking for legislation regulating this technology and imposing mandatory controls on all implementations.
  • by mi (197448) on Wednesday May 17, 2006 @07:21PM (#15354914) Homepage
    Currently the cheap and slow large-capacity harddrives are the best deal per gigabyte for data-storage. And they are always on-line. If you want an off-site backup — simply arrange with a like-minded person (or company) to exchange your (encrypted) backups via the internet. A fully automated solution without the daily trips to the safe-deposit box (which is what you are supposed to do with the tapes).

    I wonder, if the disk sizes will keep the dominance in 5 years. They probably will... Or, a major breakthrough in, say, "flash" storage technology will make all other media obsolete...

  • half the battle (Score:3, Insightful)

    by NMerriam (15122) <NMerriam@artboy.org> on Wednesday May 17, 2006 @07:36PM (#15354987) Homepage
    In my experience, the trouble is not in writing lots of data to tape, it's in reading it successfully afterwards. /only half-joking
  • That's so 20th Century....
  • by Cliff Stoll (242915) on Wednesday May 17, 2006 @07:43PM (#15355019) Homepage
    In the 1950's, data was recorded on 7-track tapes at 200 bits per inch. Those 7 tracks recorded 6-bit characters (EBCDIC?) with one bit of parity. A 2400 foot tape might hold some 5 million characters (upper case only, please). By the mid 1960's tape densities more than doubled to 556 bpi.

    Sorting algorithms were written to sort information using mag tapes; the speediest would make the tape vibrate in the vacuum columns with a minimum of reel-to-reel motion.

    By the 1970's, most shops changed to 800bpi, 9-track tapes, which would happily handle 8-track encoding. Then came 1600 bits per inch -- you could store an amazing 50 megabytes onto a single tape.

    There was a constant temptation to compress data so as to stuff as much onto the tape as possible. As a result, many graduate students earned their assistantships by decoding tapes written with oddball parity, density, and encoding combinations.

    The scattering matrices from my dissertation are encoded onto 9-track 1600 bpi tapes, carefully stored in my climate uncontrolled attic.
  • I'm sure you will need this technology to back up Duke Nukem Forever, since the Holographic RAM [prevayler.org] IBM was supposed to release in 2003 core dumps to tape.
  • Diskettes are magnetic media, just like tape is... is there any particular reason that you couldn't get the same storage density on a floppy? It'd at least have the advantage of much faster seek times than tape, I'm sure.
  • How many tapes would it take to backup Google?

"No problem is so formidable that you can't walk away from it." -- C. Schulz

Working...