Forgot your password?
typodupeerror
Data Storage Hardware Technology

Table Salt Could Help Boost HDD Storage Density By a Factor of 5 142

Posted by Soulskill
from the tasty-performance dept.
hypnosec writes "A team of researchers has managed to boost storage density on traditional magnetic platters as high as 3.3 terabits per square inch using a technique that relies on NaCl — table salt. (Comparatively, a recent 4TB Seagate drive had an areal density of 625Gb per square inch.) A research team used a technique called nanopatterning to create arrays of magnetic bits that have more regular features (PDF) than the current traditional, randomly distributed technique. Team leader Joel Yang compares the technique to a well known traveling trick; 'It's like packing your clothes in your suitcase when you travel. The neater you pack them the more you can carry.' Yang said, 'In the same way, the team of scientists has used nanopatterning to closely pack more of the miniature structures that hold information in the form of bits, per unit area.'"
This discussion has been archived. No new comments can be posted.

Table Salt Could Help Boost HDD Storage Density By a Factor of 5

Comments Filter:
  • by uigrad_2000 (398500) on Friday October 14, 2011 @12:35PM (#37715442) Homepage Journal

    It's like packing your clothes in your suitcase when you travel. The neater you pack them the more you can carry.

    Dr. Yang continued. "For speciality file systems, imagine you are travelling for a wedding, and you need to pack a suit. The extra meta data for the file system is stored in a container much like the suit compartment of your luggage."

    Unfortunately, the metaphor did not stop there.

    "Data read times have been improved also. Imagine again that the suitcase is packed neatly, but this time all clothes are on their sides. Now, imagine the suitcase is being spun in an x-ray device by the TSA. The tighter packing allows them to see more of what is packed in the suitcase during each arc of 30 degrees."

    The rest of the conversation has been edited out, but it related seek times to finding shoes that match your outfit.

    • by idontgno (624372)
      I see that yet again, Dilbert [dilbert.com] has pioneered critical storage technologies.
      • by Joce640k (829181)

        Whoah, dude. You can quote Dilbert back to 1993?

        • by Khyber (864651)

          Some of us can quote Dilbert even further than that.

          I still have something from 89 when it first came out.

      • by TheLink (130905)
        Seriously though, when I pack for travelling I roll up my clothes tightly but I don't twist them. That way they aren't really creased (in fact they are less creased than if I folded them "neatly")

        Sometimes if I need more space I stuff the rolled up clothes in a plastic bag, and use a vacuum cleaner to suck some air out of the bag (to squeeze stuff more) - but you shouldn't do this if you might not have access to a vacuum cleaner for repacking and you won't have extra space (given away gifts, used consumable
      • by Bucky24 (1943328)

        This is the infamous comic strip that offended a real-life person with the surname "Dork."

        Poor guy....

    • by MozeeToby (1163751) on Friday October 14, 2011 @12:49PM (#37715638)

      The whole time I was reading the analogy I was expecting him to finish with "That is a very graphic analogy which aids understanding wonderfully while being, strictly speaking, wrong in every possible way." With apologies to Sir Pterry.

    • Re:Metaphors (Score:5, Interesting)

      by nabsltd (1313397) on Friday October 14, 2011 @01:05PM (#37715804)

      The rest of the conversation has been edited out, but it related seek times to finding shoes that match your outfit.

      This post makes me feel like I'm reading a Douglas Adams book. Well done.

    • by Megane (129182)
      Sorry, this is Slashdot. We use automotive analogies here. Does it have anything to do with melting ice on bridges?
      • by rossdee (243626)

        Here they use salt on other parts of the road, not just bridges. (of course by the time the plow truck has got down our end of the street its just about run out of the salt/sand mixture. - and Sodium Chloride doesnt do well when its below 0F

      • Okay, so imagine you're a car dealer... if you have a parking lot full of cars parked at haphazard angles, then you won't be able to fit as many in . . . but if they're all at 30 degrees from the customer service entrance, then you'll fit at least 3 times as many cars into the parking lot and you can use a pair of binoculars to do your inventory.
  • by vencs (1937504)
    Expected the TFA to have any detail about read/write speeds - something that one would expect about a HDD (not a lame suitcase analogy). Higher densities on platters often resulted in slower IO speeds as the heads proved not to be that precise in deciding whether a bit was set or not and so ends up in verifying the data using some ECC type mechanisms. May be the real motivation is to say that disk is the new tape.
  • by Baloroth (2370816) on Friday October 14, 2011 @12:40PM (#37715516)
    Do not try this at home. Pouring table salt on your hard drive platters will not improve their storage density.
    • Re: (Score:2, Funny)

      by Anonymous Coward

      But it makes them more delicious

    • by EdZ (755139)
      It will, however, significantly improve their flavour. Unseasoned glass substrate platters are particularly unappetising.
    • You don't know that. How could you, since it will likely make them unreadable. For all you know, you may have increased the available density by more than 5X. The heads would simply not be able to read them. ;-)

      • Re:Important note: (Score:5, Insightful)

        by Jeng (926980) on Friday October 14, 2011 @12:59PM (#37715740)

        Whenever the subject comes to data density I recall Heinlein take on this.

        Basically you can take a match stick, and put a single mark on it. The distance between the mark and the end of the stick is the data being recorded. The higher precision the larger amount of data being recorded. With high enough precision a single mark could contain all the information that mankind has ever produced.

        • Basically you can take a match stick, and put a single mark on it. The distance between the mark and the end of the stick is the data being recorded. The higher precision the larger amount of data being recorded. With high enough precision a single mark could contain all the information that mankind has ever produced.

          You just re-invented arithmetic coding. Any file can be represented as a single real number between 0 and 1, as long as you have enough precision.

          • by nedlohs (1335013)

            Except that quantum mechanics don't stop arithmetic encoding from working.

            • by Artifakt (700173)

              Quantization itself most certainly does stop arithmetic encoding from achieving infinite precision, The fact that the real universe seems to have a minimum length and a minimum interval is why, for just one example, storing all the information that has fallen into a black hole requires the surface area of its event horizon. For the simpler case of a matchstick, you simply cannot infinitely divide the matchstick.
              1.616199(97)×10(E-35) meters is as small as things get. That's about 10(E-20) x the diameter

              • by nedlohs (1335013)

                Arithmetic encoding isn't marking the length of something and hence running into the planck length rather quickly. It's just a type of compression algorithm, used in practice . I can encode video with dirac just fine thanks, withtout quantum mechanics making it not work - since arithmetic encoding isn't about measuring or marking anything.

        • The information density is still limited by quantum gravity, which predicts discreteness of space-time at the Planck length scale.

        • That's not true. The number of marks you have after you write it is itself information. You'd be writing infinite data on a match stick!

          (Sorry, I've had a terrible week programming.)

        • Re:Important note: (Score:4, Interesting)

          by falzer (224563) on Friday October 14, 2011 @02:15PM (#37716602)

          If my math is right, Planck's length as your resolution limit gives you 6.187x10^34 possible marking positions per meter of stick, which means you can encode about 115 bits with one mark on a 1m Planck-grade stick.

          • Re: (Score:2, Insightful)

            by IorDMUX (870522)

            If my math is right, Planck's length as your resolution limit gives you 6.187x10^34 possible marking positions per meter of stick, which means you can encode about 115 bits with one mark on a 1m Planck-grade stick.

            Not quite. You can record one 115 bit value, which is very different from 115 bits. a 115 bit value has 2^115 = 42 million billion billion billion (10^34) bits of information.

            A slight difference.

            • You can record one 115 bit value, which is very different from 115 bits.

              No, storing one arbitrary 115-bit value is exactly equal to storing 115 bits. The relationship is one-to-one:

              8-bit value: 10110110 (181/255 ~= 0.713 units)
              8 bits: 1 0 1 1 0 1 1 0

            • by artor3 (1344997)

              No, he had it right. There are ~6(10^34) = 2^115 marking positions. You mark one of them. That means you can select one value out of 2^115.

              If I let you pick one value out of 2^10 = 1024, would you say that you have 10 bits of data or 1024 bits of data?

              Or, to make it even easier, if I let you pick one value out of 2^0 = 1 (as in, you have to pick the number one and have no other option), is that zero bits of data or one?

              • by IorDMUX (870522)

                If I let you pick one value out of 2^10 = 1024, would you say that you have 10 bits of data or 1024 bits of data?

                Well, actually, in my world of mixed signal and ADC design, we would refer to the resolution of the data as 10-bit. i.e. a 10 bit value. That 10 bit value can represent 1024 unique possibilities, or 1024 least-significant-bits (or, as we say, "bits") of data. For example, if our 10-bit ADC has 10 bits of error, then it is only a 1 % error.

                Maybe it's an industry nomenclature thing. *shrug*

                • You're confused. 1024 possibilities is exactly the same as 10 yes/no questions. It takes 10 questions to find one value among 1024, using binary search.

                  So even though 2^115 is a big decimal number, that number represents only 115 independent yes/no questions. Or put another way: if you have 115 yes/no answers, and you want to represent them as a single symbol, then you'll need a huge number of symbols to handle all the possibilities, 2^115 different symbols, in fact.

                  • by IorDMUX (870522)
                    I believe I understand what you are saying... After all, your binary search is my successive approximation ADC [wikipedia.org]

                    But I don't see how that disagrees with my previous point.
                    • In your previous points on this thread, you've repeatedly used "bits" incorrectly. A bit [wikipedia.org] is the smallest unit of information. It represents an atomic yes/no question, a single truth value in logic, etc. When you have 2^10 different values, those 1024 values aren't bits. 1024 is only the size of the range of a 10-bit number.

                      In your ADC example, the bits are the actual 0/1s in the approximation, but the bits are not the waveform approximation itself. That's an actual signal intended to recreate the input s

      • You don't know that. How could you, since it will likely make them unreadable. For all you know, you may have increased the available density by more than 5X. The heads would simply not be able to read them. ;-)

        I just tried that with my PS3, and I think your theory may be correct. It doesn't turn on, but I get a sense that the density has increased by roughly 500%. I also get a sense that I need to buy a new PS3. I'll definitely try to transfer the HDD over to the new one!

    • Re:Important note: (Score:4, Interesting)

      by interval1066 (668936) on Friday October 14, 2011 @01:27PM (#37716076) Homepage Journal
      Took me a minute becuase the article was very difficult for me to read (navy font against a lighter blue background, brilliant) but all they did was add table salt (in aqueous solution I'm sure) to the developing chemicals in the etching process. That's interesting from a chemistry persepective; it implies they may be able to even higher densities by fiddling around with that catalyst- in stead of salt maybe they add shrimp cocktail to the process and get peta byte capacities or something.
    • by drinkypoo (153816)

      Do not try this at home. Pouring table salt on your hard drive platters will not improve their storage density.

      That depends on which definition of the word "areal" you use.

  • Salt is now good for both water AND data rentention YAY!
  • I haven't read the article, but the idea sounds kosher.

    Good night, tip your waitresses.
    • by tgd (2822)

      I haven't read the article, but the idea sounds kosher.

      Good night, tip your waitresses.

      I'm taking the whole thing with a grain of ...

      Nevermind.

  • ...does it make data storage more palatable, or should the claim be taken with a pinch of salt... :)

  • We should take that news with a grain of salt. One can not just spread NaCl on its hard drive and get performance kick.

  • I want a massive SSD capacity increase, and price drop.

    • by ajlitt (19055)

      Until that happens, enjoy the increasingly more garish whips the buggy drivers will be brandishing.

    • I would enjoy this as well, but I fear that we're never going to see SSDs that contain more storage for less cost.

      While platters are often pretty pricy glass, it's still not up there with high purity silicon wafers. In addition, you only have to deposit an even layer of magnetic material - with flash you need not just the semiconductor gates, but the paths to them.

      As such, I think that hybrid drives, such as released by Seagate, will eventually dominate. For desktops, hard drives are plenty fast enough fo

    • Well at least the big price drop part. Ignoring the specifics its like wishing for Ferrari's to drop down to Accord prices. Both products perform the same basic purpose but one uses much different engineering to accomplish the same task. You can make the Ferrari engine cheaper by mass producing it, but it's just always going to be really expensive to produce no matter what. Same thing with SSDs.

      Oh how I long for cheap 1TB SSD drives and Gigabyte ethernet Internet wide...

  • Just don't use too much, or your disks might begin retaining water...
  • This probably won't happen, but:

    If it drives up the price of salt then it may spur desalination projects making more drinkable water available. It might make desalination cheaper, and help increase the world's water supply. However, you'd need to use truly huge quantities of salt for that to happen.
    • by Anonymous Coward

      Have you any idea of the cost of salt these days? It's almost free.

    • As salt is one of the most plentiful minerals on the planet and such a small amount of the overall salt produced is due to desalinization, I doubt your prediction will ever happen.
  • Whats the chemical formula for driveway salt? Kosher salt? Sea salt?

    Imagine how much we could store in the big granules of road salt when winter rolls around!

  • Would it have killed you to call it sodium chloride in your summary and skip over the table salt thing?

    • Table salt has a 1:1 correlation to sodium chloride. Therefore no additional information is conveyed using the chemical name. While I appreciate your desire to be geeky, may I point out that "table salt" is 33% more efficient at conveying the intended information?

      • by epine (68316)

        While I appreciate your desire to be geeky, may I point out that "table salt" is 33% more efficient at conveying the intended information?

        Tell that to my Himalayan salt or my Mediterranean sea salt, both of which imply unspecified exotic trace elements. I'm not quite so willing to energize on white bread custom as to equate the two. For brevity, I keep a shaker of NaCl.

        • by AdamHaun (43173)

          Tell that to my Himalayan salt or my Mediterranean sea salt, both of which imply unspecified exotic trace elements.

          They're the same trace elements as all unprocessed salt -- minerals that were also dissolved in the ocean. I ran across a site that claimed Himalayan salt has 84 elements, although that's impossible without including some toxic and/or radioactive ones. The claim seems to be based on a lab report that lists 84 elements, many of which are not present in detectable quantities in the salt.

          [I was in

      • Table salt is inappropriate for almost all chemical processes because of additives and impurities.

      • by afabbro (33948)

        Table salt has a 1:1 correlation to sodium chloride.

        Some of us like to get some potassium [mortonsalt.com] in our "table salt".

  • by Anubis IV (1279820) on Friday October 14, 2011 @01:02PM (#37715776)

    This is one story that I'll be taking with a grain of salt.

    buh-dum-TISH

  • Hey fellas, it's us. You can tell us that one of you was having his lunch at his desk and spilled some of the salt from his salt packets onto some of your test disks. And actually tried to pass it off as 'more experiments'. He could have put the devices out of the way but naaaa. Anyway, win-win, right? (Oh, could you hand me some of the pepper, while you're publishing your results...)

  • Does the non-repeating nature of quasicrystals help (or hurt) data storage?

    Any Nobel Laureates care to reply?

  • All I'm sayin' is, take it with a pinch of salt.
  • Salt + Oxygen + Iron Oxide (rust).

    How long will this chemical combination remain stable? Is long-term oxidation a concern here?
  • That mark is useless without the table that tells you what each distance represents. So on that matchstick you haven't stored data, merely a hash of that data.

  • by idontgno (624372) on Friday October 14, 2011 @04:21PM (#37717898) Journal

    if you use the platters to cook your shredded potatoes for breakfast while computing and storing cryptographic trapdoor values, you'll discover...

    <sunglasses>

    your hashes are already salted.

    YEEEEAAAAAAAHHH

  • Don't forget, hard disk density is inversely proportional to reliability and life expectancy. I still use 40-80 gig drives that are over 10 years old for all my os drives and you know what? I never suffer data loss. This is because on a low density drive you have a significant amount of redundancy in the wear and tear of moving parts before the errors become unrecoverable. The drives basically get slower and slower as errors increase read time, but they don't fail. This leaves plenty of time to replace the
  • Honestly why does no one seem to notice the part about this where the salt is being used in a "photo" resist for Electron Beam Lithography. That's beam - as in every single surface feature needs to be drawn by a beam of electrons one at a time. The amount of time and expense that would go into the construction of even one 3.5" platter is staggering. Yes it's cool and all but only a military application or James Bond could justify it; ever. An improvement to a mass-produced technology that makes it impossibl

E = MC ** 2 +- 3db

Working...