Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

The Hard Drive Turns 50

Posted by samzenpus on Wed Sep 13, 2006 09:55 PM
from the but-it-still-feels-10 dept.
JHU writes "When the hard drive was first introduced on September 13, 1956, it required a humongous housing and 50 24-inch platters to store 1/2400 as much data as can be fit on today's largest capacity 1-inch hard drives. Back then, the small team at IBM's San Jose-based lab was seeking a way to replace tape with a storage mechanism that allowed for more-efficient random access to data. The question was, how to bring random-access storage to business computing?"
+ -
story

Related Stories

[+] 50th Anniversary of the First Hard Drive 225 comments
ennuiner writes "Over at Newsweek Steven Levy has a column commemorating IBM's introduction of the first hard drive 50 years ago. The drive was the size of two refrigerators, weighed a ton, and had a vast 5MB capacity. They also discuss the future of data storage." From the article: "Experts agree that the amazing gains in storage density at low cost will continue for at least the next couple of decades, allowing cheap peta-bytes (millions of gigabytes) of storage to corporations and terabytes (thousands of gigs) to the home. Meanwhile, drives with mere hundreds of gigabytes will be small enough to wear as jewelry."
This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More
Loading... please wait.
  • by Anonymous Coward on Wednesday September 13 2006, @09:59PM (#16101263)
    Has anyone run HD Tach on that original IBM hard drive?
  • I seem to recall reading this story TWICE before this one!

    I've found one of them: [url:http://hardware.slashdot.org/article.pl?sid=0 6/07/30/2124225], but I KNOW there was a second one.
  • by red_crayon (202742) on Wednesday September 13 2006, @10:01PM (#16101268)
    I used a hard drive when they were the size of a suitcase.

    That's nothing. I used a hard drive when they were the size of a VW and held only 64 bytes. That's bytes not kb.
  • by Anonymous Coward
    I've got a few in my flying car.
  • I predict (Score:5, Interesting)

    by grasshoppa (657393) <(skennedy) (at) (tpno-co.org)> on Wednesday September 13 2006, @10:07PM (#16101292) Homepage
    At some point in the future, capacity will take a back seat to recoverability ( for the average consumer ). To that end, I predict harddrive companies effectively setting up a raid 1 array on a single drive; Probably by platter. To the host system, it would appear as a single drive of 160gb ( for example ), but it would actually be two platters of 160gb, with a bit for bit copy being maintained on the fly by the drive itself.

    Access would be through a standard API.

    Extending this further, we could add even more intelligence to the drives, and with the sacrifice of more storage space, would could have the drive taking care of shadow copies ( this operating under the assumption that the host system knows how to handle the drive ).

    This is the direction I predict for future harddrives; At some point we will come to a place where we don't really need the extra capacity. At that point the harddrive manufactures will begin to add more intelligence to the drives.
    • Re: (Score:3, Informative)

      We're already on the way with SMART, with many (most?) drives having reserved sectors that get mapped over bad sectors when they crop up. This won't be able to recover lost data, but a drive that verifies writes could re-write a sector that didn't make it to disc on first attempt.

    • Re:I predict (Score:5, Insightful)

      by nmb3000 (741169) <nmb3000@that-google-mail-site.com> on Wednesday September 13 2006, @10:20PM (#16101352) Homepage Journal
      [automatic internal redundancy]

      The problem I see with this is that (in my experience) there are several single points of failure in a hard drive, and if one of them goes the entire drive is toast. Specifically, the heads, the motor, and the controller board. I've had all three die on different occasions, and for all three the entire drive is dead. If the motor or controller board fails, then your data is fine, but you'll need to spend up to $1,000 (or more) to get the data off the drive. If the heads fail (mechanically or physically) there is a good chance that all the platters can be damaged so you're totally screwed.

      In any case, aside from tons of bad sectors forming on the drive (in which case the entire drive is probably on it's way out) I don't see how an internal mirror can help much. You can't recover the data without going through an expensive data recovery service, so you may as well just buy a second physical drive, something that anyone can swap out and replace.
      • I must agree with this, when I first read the grand-parent I was thinking, whats the point of a raid if it's all on the same drive? I mean technically I could do that right now with linux, just make 2 partitions of equal size and software raid across them. Pretty pointless....now maybe if it had 2 separate little mini-drives inside of one that might technically count....not entirely unfeasible if platter density increased enough, but still this seems a little silly to me...
      • Actually, not so with the controller board. If you buy a drive of the exact same model (hopefully of the same batch - best yet, buy an extra initially and have a spare), you can remove the controller board from a working drive and replace it onto the drive with the bad board. There's almost never actual solder points connecting the board to the drive itself, but rather an array of pins and sockets that should separate with relative ease. Assuming that the board frying didn't damage anything internally (g
    • Re: (Score:2, Interesting)

      Interesting point.

      Ok fellow geeks. What are everyones' predictions about what computer storage will be like in 50 years? Include capacity, medium, and whatever else you want.

      My guess is with organic/biological storage with essentially unlimited capacity - if you need more just grow more.
    • Re: (Score:3, Interesting)

      To that end, I predict harddrive companies effectively setting up a raid 1 array on a single drive; Probably by platter.

      Why would that be at all useful?

      On my mental list of potential failure points, damage to the platter doesn't rank very high.

      Other than the occassional bad sector, if you're going to get data corruption (or physical damage), your data is going to get FUBARed on both platters.

      I agree with your conclusion about more intelligence, just not the notion that a one-drive RAID-1 would make any sens

      • Re: (Score:3, Interesting)

        Personally, I'm waiting for them to cram 2 opposing sets of read/write arms (or even just a second set for reading) so that they can effectively halve the latency and seek times without having to go faster than the existing 15k screamers.

        For a short time Seagate made a series of drives with dual head assemblies for transactional processing but they were not cost effective. I do not remember how the interfacing worked.
    • At some point in the future, capacity will take a back seat to recoverability ( for the average consumer ). To that end, I predict harddrive companies effectively setting up a raid 1 array on a single drive; Probably by platter. To the host system, it would appear as a single drive of 160gb ( for example ), but it would actually be two platters of 160gb, with a bit for bit copy being maintained on the fly by the drive itself.

      That's not going to help you if the motors and/or onboard circuitry dies. Which i

    • Re: (Score:3, Insightful)

      I know it's not happening this year or next year, but I would expect that the real solution is going to be moving away from hard discs for storing important data.

      It may be a decade before most people switch to some form of non volatile memory for new purchases, but I would expect it to be reliable enough, and hopefully by then issues of Windows writing too often to drive will be fixed, as well as hopefully eliminating the need for a swap file.

      This is all hypothetical. It's hard to know for sure when hard d
      • Re: (Score:2, Informative)

        It may be a decade before most people switch to some form of non volatile memory for new purchases, but I would expect it to be reliable enough

        Yes, but still not impervious to Rhinoceros attacks, which are very very common where I'm from (Florida). And what about when the glaciers slide off Greenland all at once and cause the 300ft tsunami all around the world, then where will your data be? Underwater, that's where.

        The greenland thing is actually possibly going to happen, as the water pools on top of the
    • Capacity is the main selling point of hard drives. When capacity stops meaning as much, people will move to things like flash for the lower power consumption.

      My prediction would be that in 10-15 years, consumer machines don't have hard drives at all as solid-state memory achieves terabyte sizes and the number of rewrites ceases to be an issue.
    • raid-1 on two platters make no sense.

      My guess would be that failure of physical platters or read-heads account for perhaps 10% of all hard-disc crashes.

      Having two platters with the same data will do nothing for you when the drive-electronics die. When the motor driving the spindle has a problem, when the stepper is no longer able to align the read-heads properly.

  • by Brickwall (985910) on Wednesday September 13 2006, @10:08PM (#16101298)
    While a student at the University of Toronto in the late 1970's, my fraternity (mostly engineers) invited a professor for a dinner. We retired to the library afterwards with a case of beer, and I ventured the comment "Won't it be great when you can get a desktop computer with 1 Mb of RAM, and a 10 Mb hard drive?".

    The prof thought this was the funniest thing he'd ever heard. He listed the following "fundamental physics" reasons why these devices would be impossible:

    1. You could never make the magnetic domains small enough to get that density

    2. Even if you could, you could never make stepper motors precise enough to read the data.

    3. Even if you could, you could never make read/write heads sensitive enough to read such small domains.

    4. Even if you could, you could never make a disk which rotated stably enough to prevent head crashes.

    5. As for the RAM, he said we could never make chip densities high enough to get 1 MB on a desktop.

    6. Even if you could, the heat generated by those RAM chips would require a small refrigerator.

    7. And finally, even if you could make the transistors small enough, you would get so many tunneling errors that the RAM would be completely unreliable.

    I wonder if he's seen an Ipod Nano yet...

    • by merreborn (853723) on Wednesday September 13 2006, @11:30PM (#16101603) Homepage Journal
      A friend's grandfather actually worked at the San Jose IBM lab back in the days when they were working on early drives -- I think he just turned 88 this month.

      At any rate, he talked my ear off for an hour once, talking about how they'd spent a bunch of time trying to figure out the optimal height above the platter to float the head at. He said they used a jet of compressed air under the head to float it, not unlike an air hockey puck.

      Long story short, if they really were working on these things in this scale back in those days, I can't say I can blame your professor -- you might as well have been talking about flying cars and having an entire meal in a single pill. I mean, hell, drives these days hold millions of times more data than they did just a couple of decades ago. I don't think anything's ever miniaturized that fast.
    • While your professor friend was being a fool Richard Feynman was writing "THere is plenty of room on the bottom". See if you can find his paper. He predicited densities much higher.

    • by Kjella (173770) on Thursday September 14 2006, @05:16AM (#16102700) Homepage
      Alright, so he wasn't a visionary but I think you can point to most computer scientists in the 1940s-1970s and laugh
      "You didn't have a clue how far computers would go".

      Then you can point to most computer scientists in the 1980s and laugh "You didn't have a clue how far Internet would go".

      Then you can point to most computer scientists in the 1990s and laugh "You didn't have a clue how far wireless connections would go".

      Then you can point to most computer scientists in the 2000s and lau... oh wait, that's us. I'm not exactly sure what they'll be laughing about, put I'm pretty sure they will. It's really easy to mock technological predictions with 20-20 hindsight. Now if you'll excuse me, I'm going for a trip in my flying car driven by cold fusion...
      • FYI, Flash memory IS Random Access Memory, and the nanos have way more than 1MB of it.
  • by sporkme (983186) * on Wednesday September 13 2006, @10:11PM (#16101310) Homepage
    My father talks about his younger days with the US Air Force as a mid-level computer technology worker in Anchorage. He speaks of how dangerous magnetic storage was in the early days, with all that weight in a drum, spinning up to 1200 RPM. We still jokes about the emergency procedures in the event of a catastrophic mechanical failure of operating storage media. The USAF's official line was to take cover in a corner behind other heavy equipment at the first sign of trouble. Techs used to work under constant threat of going three rounds with bouncing betty. Now all we have to worry about are laptop batteries.

    See Drum Memory [wikipedia.org]

  • Code name was Ethel (Score:3, Informative)

    by BadAnalogyGuy (945258) <BadAnalogyGuy@gmail.com> on Wednesday September 13 2006, @10:11PM (#16101313)
    It's kind of a strange coincidence that the codename for the hard drive project was Ethel because the same day that these huge hunks of iron were debuted was the day that Hurrican Ethel formed in the Gulf of Mexico (it made landfall the next day in Mississippi).

    These days we're talking about capacities that can hold all the information of every hurricane evar on a single disk. What a ways we've come.
    • "These days we're talking about capacities that can hold ALL the information of every hurricane evar on a single disk. What a ways we've come."
      (emphasis mine)

      I'll be pedantic.

      You're not thinking big enough. ALL of the information would be the location of every molecule of air, etc at every point in time during the hurricane. For that, we would need a hard drive as massive as the hurricane for each point in time. I think we would quickly run out of mass in the universe if we stored ALL of the information.
  • Big bits (Score:3, Interesting)

    by NMBob (772954) on Wednesday September 13 2006, @10:37PM (#16101421)
    When I was in high school (1970's) our computer programming/math teacher had a hard drive disk platter that might have been from one of the these machines. I seem to recall that it was larger than 24" in diameter, but maybe I was just smaller. Anyway, the disk had some silver powder on it -- magnetic I'd guess -- and you could actually see the individual bits. They were pretty thin, but the tracks looked to be about 1/8" wide/tall.
  • by Easy2RememberNick (179395) on Wednesday September 13 2006, @10:49PM (#16101463)
    At 50 years old I bet it's more floppy drive than hard drive.
  • by sokoban (142301) on Wednesday September 13 2006, @10:58PM (#16101497) Homepage
    "1/2400 as much data as can be fit on today's largest capacity 1-inch hard drives"

    Really now, that is almost completely uninformative since most people have no idea what the capacity is of today's largest 1 inch hard drive. I know that it is cool and all how much storage has shrunk, but I think just saying 8 megs (or whatever the storage capacity was) tells people more than saying a fraction of an obscure unit.
    • "1/2400 as much data as can be fit on today's largest capacity 1-inch hard drives"

      Really now, that is almost completely uninformative since most people have no idea what the capacity is of today's largest 1 inch hard drive. I know that it is cool and all how much storage has shrunk, but I think just saying 8 megs (or whatever the storage capacity was) tells people more than saying a fraction of an obscure unit.

      Yeah might as well go for the "sidewalks to the moon" or "statue of liberty on is side" compa

  • how to bring random-access storage to business computing?

    Now, the question is how to best make use of the *non*-random-access storage that business computing has available? Most people think of hard disks as random access, but really they're not -- there's a huge performance penalty for random reads and writes. A disk that can do many tens of MB/s of sequential reads can only do maybe 200 4kB sector reads per second. That's a *huge* difference. So much so, that it's almost free to just read a bunch o

    • Same thing with cache operation as well as the latest double-data-rate RAM. Cache is transfered a block at a time, to reduce overall latency, and they do it because most of the time you read and/or write multiple data elements within a small contiguous memory area, then move on. Even with small caches that are used in embedded CPUs you end up with cache hit rates that exceed 90 percent for most common applications.
  • I remember working with fixed hard drives (i.e. non-removable) that were 500 MB, and larger than washing machines.

    I remember having colleagues who broke their feet after removable hard drives fell on them (those were only 200 MB, but HUGE ...

    The same place I worked at had XT like PCs with external hard drives in shoe box sized housing.

    Those were from the mid-80s by the way ...
  • by kyjl (965702) on Wednesday September 13 2006, @11:52PM (#16101675)
    September 14th, 1956: The first time porn is loaded onto a Hard Drive