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

 


Forgot your password?
Close
typodupeerror
×
Data Storage IBM

Fujifilm and IBM Set World Record With 580TB Magnetic Tapes (pcmag.com) 53

Posted by BeauHD from the new-and-improved dept.
An anonymous reader quotes a report from PCMag.com: Mark Lantz, Manager of Advanced Tape Technologies at IBM Research, explains how researchers at IBM and Fujifilm brought together more than 15 years of work to set a new world record in tape storage. What they achieved is an areal density of 317Gbpsi (gigabits per square inch), which translates to a single tape capable of storing 580 terabytes of data. In order to achieve such a high areal density, the research team had to develop a brand new tape and created Strontium Ferrite (SrFe) in the process. Existing magnetic tapes rely on Barium Ferrite (BaFe), but SrFe offers the potential for higher density storage in the same amount of tape. Alongside that, the team also "developed a family of new servo-mechanical technologies including a new servo pattern that is pre-recorded in the servo tracks, a prototype head actuator and a set of servo controllers." The end result is a very high capacity tape that can be read while moving at a speed of 15km/h (9.3mph).
This discussion has been archived. No new comments can be posted.

Fujifilm and IBM Set World Record With 580TB Magnetic Tapes More

Fujifilm and IBM Set World Record With 580TB Magnetic Tapes

Comments Filter:

  • ... you won't be able to find a drive to read them. Oh well.

    • When tape storage was mainstream you really did not see a lot of those rare earth magnets. Now they are everywhere. Imagine the hell of some desk fidget toy ruining a set of backups.

    • Re: (Score:3)

      by kenh ( 9056 )

      The IRS maintains a facility to read old tape formats, so that businesses can't claim their data is "unreadable" during an audit or investigation.

      • Re: (Score:2)

        by AvitarX ( 172628 )
        In the course of normal business the IRS can only go 5 or so years I thought.

        Maybe they want to go back a few more to get a scope of things, but 30 seems a stretch.

        They can go back further specifically for fraud, but again, 30 years? seems unlikely.

        Also, like, oops, nobody knows where the encryption key is or the password (though I guess 30 year old crypto tech may not actually keep things secret).

        • Re:And in 30 years... (Score:4, Informative)

          by Joce640k ( 829181 ) on Friday December 18, 2020 @09:35PM (#60847116) Homepage

          In the course of normal business the IRS can only go 5 or so years I thought.

          They can go back as far as they want if they any fraud in those 5 years.

          • Re: (Score:2)

            by AvitarX ( 172628 )
            Sure, but there's diminishing returns in storing equipment and the equipment to read it.

            Especially if the idea is to be "got yeh" when backups are encrypted and companies are under no duty to maintain access to old records anyway.

            I don't doubt that they can go back 10 years, maybe even 20 depending on the popularity of the medium, but 30 years seems a stretch.
      • Tapes are not always stable. Some acetate tapes from the late 1950s and early 1960s did not age well at all. The coating binder deteriorated, and you could wipe the tape down to clear base between two fingers, without much pressure.

        • Tapes are not always stable. Some acetate tapes from the late 1950s and early 1960s did not age well at all. The coating binder deteriorated, and you could wipe the tape down to clear base between two fingers, without much pressure.

          Actually, the Acetate tapes from the â50s and early â60s are more stable. It is the Polyester (Mylar)-based tapes of the late '60s and '70s that have the issue with the binder.

          Of course, you can usually keep them alive for one more playing (to transfer to a newer format) by baking them in a 150 deg F oven for a couple of hours before playing them for the first time.

      • Re: (Score:2)

        by hawk ( 1151 )

        yet they can't find their *own* data from a decade or so ago . . .

        (seriously, folks, keep a copy of your tax return in case you ever need it down the road. Perhaps the latest electronic system will keep what's filed now available indefinitely, but if you need to prove to the IRS what you filed ten years ago, they usually can't come up with their copy or a facsimile . . .)

        hawk

    • Well, 30 years ago you used to be able to buy Exabyte tapes while these new tapes are only 1/2000th of an exabyte.

    • You aren't wrong.

      I fear for the future of data archiving. Folks see "tape" and assume they'll be able to access said data 10+ years in the future, never thinking about the "how". Data archiving is not a passive activity; you need to actively keep the data in a readable format, on a medium which is usable. You can't just fire and forget with this stuff; we're not there yet ( ironically supported by this very announcement ).

      We'll get there some day. Maybe in 50 years we'll be ready for an "archival format

      • For this very reason, it almost seems as for archival it would be best to just keep the data in an on-line storage. Build up a storage array that will meet your foreseeable needs for about 5 years. Then in 5 years, build a new storage array double the size or larger if your data needs have grown, migrate all the data and rinse and repeat every 5 or so years. These extremely large tapes could be used for offline shorter term storage. You know in the case that your entire on-line array gets infected by some r

    • In 30 years, the drive that is supposed to read them won't be able to read them either.

      A disk lacks the problems such as print through, stretching (to the extreme tapes go though) and other issues in regards to tape storage.

      Yeah, they packed a ton of data on very tiny areas of magnetic tape, but I wouldn't expect this to remain error free even one year from now. This was just a tech demo. Real world usage would mean larger magnetic areas pr bit, redundant copies, and error correction which will bring the st

    • Comment removed based on user account deletion
  • We know now on what medium Windows 11 will be delivered

    • Yeah that'll be cartridge #1, and the HP printer driver with the adware/spyware/nagware/eyecandy will be cartridges 2 and 3.

  • As we all know a backup isn't a backup until you have successfully done a recovery. I don't know about you but I have had instances where the backup was supposedly completed in an enterprise environment and when an event required a recovery, well let's say it did not go well. Be interesting to see how redundancy/testing is done to ensure a good backup with such large tapes

    • If you want to be really sure then you must punch your data onto paper table using a three ring binder hole punch, or a grommet punch will do in a pinch.

      • s/stable/tape/ and screw slashdot for not supporting post redo. Not getting enough traffic to pay for a competent coder or what?

        • Not getting enough traffic to pay for a competent coder or what?

          No it's so that you can't change the meaning of the post after to make the replies look bad.

          • Not getting enough traffic to pay for a competent coder or what?

            No it's so that you can't change the meaning of the post after to make the replies look bad.

            That theory is fucked up, does not improve the quality of posts at all, in fact makes them worse, and drives users away. (e.g., me. See you)

            Proper way to achieve the effect you tout is, allow redo up to some number of seconds after original post. If you must enforce this, which has been proved (by reddit) to be unnecessary anyway. I return to the theory that slashdots can't code.

            • Slashdot is lots older than reddit. It was a sort of experiment that ended up how it did. This includes some weird traditions that give it a specific character including not allowing editing and "never" deleting posts. The latter has been compromised, so you can now DCMA notice a post and stuff, but all the N-a-z-i spam stays up for a reason. This pushes up the importance of the moderation system. I see missing posts on Reddit and they can actually be a problem. It's just something to accept and be a bit

      • Better make sure your paper is acid-free then.
        And not tasty to tiny animals either.

    • Re: (Score:3)

      by dshk ( 838175 )
      Tape backup is the most redundant backup by its nature. You change tapes (or a tape library change them) typically at least once a day. Moreover, most tapes are offline or write-once and some are usually offsite. The read head is behind the write head, so the tape drive immediately reads back anything which is written out and if there is an error it fixes it. In addition ecc is used if some bits go wrong later.

    • do you only do one backup? tapes like these you do a daily full and then log or incremental during the day. the entire tape has to die in the robot for all the restores too fail. disk can be destroyed and wiped

  • In 21st century units (Score:3)

    by Tough Love ( 215404 ) on Friday December 18, 2020 @07:43PM (#60846870)

    For those of us who live in the 21st century, that is almost 50 gigabits per square centimeter, or about .5 gbit/mm^2.

    • Look, trying to get the US to switch by pointing out how old their unit system doesn't seem to be working so we need to try a stealthy approach. Don't go all the way to gbit/cm^2, meet them halfway with 125 Gbit/in/cm. Next, we get them to adopt the kilo-yard instead of mile for distance, and pretty soon we'll have them on the SI system without them ever noticing.

      • Hahah...well "old" ? how about "shitty", illogical historical histerical units here, eh? Event the stagnant, fogged up England changed (well almost)...

        https://en.wikipedia.org/wiki/... [wikipedia.org]

        Seems like US always goes for short time gain so big, long-term benefit changes are "freedom killer" events.

        • actually, when you actually look at them, the breakdowns are *quite* logical.

          A foot, for example, can easily be split into halves, thirds, quarters, or sixths still measured in common units.

          Larger units were made from useful multipliers of smaller units, not an arbitrary assignment of ten.

          There really wouldn't be much benefit to change from one system of uniform units that people already know to another. Folks have been trying to make us do this since the *Jefferson administration.*

          It ws a different story

  • 580TB...if you can wait nine years (Score:5, Interesting)

    by prisoner-of-enigma ( 535770 ) on Friday December 18, 2020 @07:55PM (#60846904) Homepage

    Being an old fan of tape, I saw this announcement elsewhere and eagerly read it. Two major takeaways that really throw a lot of cold water on things:

    1. 580TB assumes 2:1 data compression. It's getting harder to find any file format these days (of any reasonable size) that isn't already compressed in some manner. The biggest stuff you can find -- video -- is almost always already compressed via H.264 or (more recently) HEVC/H.265. You don't get diddly squat for compression on a file that's already compressed. That 2:1 figure is going to be hard to reach.

    2. Despite what the summary tries to imply, this 580TB tape cartridge does not actually exist. Sure, they have the areal density required to do it...in a lab. I immediately wondered how long it would take to get this out of a lab and into a working product. The source article happily had that information: around 2030. So, around nine years from now you might be able to buy a product that might realistically hold 300TB, maybe 400TB if you're lucky.

    Is it wonderful achievement to put nearly 300TB (uncompressed) on a tape today? Sure. Nine years from now it'll look a lot less impressive. It'll probably look more like the problem we have with tape today, namely that data volume grew just as fast -- if not faster -- than tape capacity. Go back nine years and look at the size of your average hard drive. Now look at what's in your rig today. I had 500GB drives then. I have 12TB drives now. That's a growth factor of 24x. To match than in tape growth, we'd need the current 12TB (uncompressed) LTO-8 to scale to 288TB (uncompressed). Which, oddly enough is almost exactly what this proposed new tape technology will give!

  • "Never underestimate the bandwidth of a station wagon filled with tapes."

    Just as true today as it was in 1981 when Andrew Tanenbaum first uttered it...

    • Re: (Score:2)

      by kackle ( 910159 )
      And as I had mentioned to @93 Escort Wagon, for better latency, fill up the gas tank before you begin "transmission".

  • When density goes up the data retention (error rate) ususally goes up too. From perosnal expeirence of running picture frame made of old laptops - the SD cards seem to start getting bit error in 3+ years. How about this storage?

  • to write one of these things ? I did RTA but did not see any mention. It mentions a density of 317Gbpsi (gigabits per square inch) and speed of 15km/h. It does not say how wide the tape (or a track) is or what the read/write speed is - so a somewhat useless article.

  • Sounds like they've got a lock on the 8k VCR market.

  • Can you imagine tape running at 15km/h? I can not.

    Let's convert the units to something more intuitive:
    15 km/h = 4.1667 m/s = 416.67 cm/s = 164.04 inch/s

    Wow, that's fast.

    • I seem to remember IBM launched the 3420 model 8 - a 200ips drive - in 1973. They were not cheap, and a lot bigger than a "full height" LTO.

      I have written 9 track tapes and then read them back 30 years later. You always keep three copies of the data in each of three physical locations if the data is worth having. Rotate them with the Grandfather/Father/Son routine. (A lot easier with LTO).

      If your database data is important, you can do write-ahead logging to a remote tape drive (or more than one) in additi

  • Now that is a lot of high quality porn! Nice!

Slashdot Top Deals

You can tune a piano, but you can't tuna fish. You can tune a filesystem, but you can't tuna fish. -- from the tunefs(8) man page

Close