Seagate ships first 8TB hard drive

Just a few months after rolling out its first 6TB hard drive, Seagate has kicked up the capacity by 33%. The firm announced this morning that it's now providing 8TB drives to "select customers." These drives are the "world's first" to hit that capacity milestone, according to marketing VP Scott Horn, and they won't be restricted to an exclusive audience for long. Broader availability is scheduled for next quarter.

The 8TB monster uses a standard 3.5" form factor and 6Gbps SATA interface. It's targeted at datacenters, cloud-based services, and bulk data storage—and it has the rotational vibration tolerance typical of drives designed to slot into tightly packed servers. That's pretty much the extent of what Seagate is revealing about the drive right now, though. The company declined to clarify the platter count, spindle speed, and whether the 8TB unit uses shingled magnetic recording technology. We were told that more details are due "in the coming weeks."

The 6TB drive Seagate announced earlier this year doesn't use SMR, but it's hard to imagine a quick jump to 8TB without it. Unlike traditional recording tech, which puts spaces between individual tracks, SMR layers them on top of each other in a staggered fashion, much like the shingles on a roof. This overlap increases the bit density of the drive platter, but it can also slow write performance, since altering data requires that overlapping tracks be read and rewritten.

Last year, Seagate said that it had already shipped "over one million drives" based on SMR. We haven't heard much about the technology since, but the new 8TB drive seems like a prime candidate for it.

Comments closed
    • Coyote_ar
    • 5 years ago

    i wonder why they dont go back to the 5.25″ hard drives for stuff thats mostly sequential reads, not seek time intensive.

    id love to have a 10+ tb drive for my movies, and other non intensive stuff.

    bring back the quantum bigfoots!!!

    with current data density, they could increase storage capacity by 250%. or even use older technology, and provide huge storage capacity for a lower cost.

      • Spunjji
      • 5 years ago

      I don’t think it would actually reduce the costs – getting platters of that physical size to move at a decent speed, reliably and without vast power draw would be non-trivial.

      Also most servers don’t have the space for drives that size now, and servers are definitely where drives like this are meant to go.

        • Coyote_ar
        • 5 years ago

        why just servers? servers need higher performance, specially in the seek time department due to multiple requests.
        plus on the server enviroments, they already have tape storage for storing stuff that you dont access often. and as you said, the format is not server friendly either (this one could be debatable, but lets say it doesnt).

        a 5.25″ drive would make sense, for example replacing those “green” drives. the performance on those disks its already shit, having a 5.25″ hdd wouldnt make a big difference. but you could have almost 3 times the storage capacity, while keeping that shitty performance.

        even the sustained transfer speed would be pretty nice, since the radial velocity on the outer edges of a 5.25″ drive is higher than a smaller, higher rpm 3.5″ drive. so you could even pull out similar performance to a 5400rpm drive, with a slower one (old quantum TS drives used 4000rpm). again … seek time would be shit, thats granted. but as long as its good enough to store those movies, music, etc … who cares?

        btw, the energy consuption of a 5.25″ drive is actually better than the performance of a similary sized and with same technology 3.5″ drive. you need lots of platters on the 3.5″ to match a single 5.25″. and WAY better if you need multiple 3.5″ drives to match the same storage. just to put it into perspective, a single 4 platter 5.25″ drive would have be able to store as much as 3-4 3.5″ drives. and keep in mind that given the format of the 5.25″ bays, you could stack tons of platters in there. 6-8 platters should be possible, thats stupid ammounts of storage on a single drive. again … performance on anything other than sustained transfers = shit.

        tbh the only concern would be the manufacturing process. i ignore if the current fabrication process would allow to manufacture 5.25″ platters.

        at the time, the quantum bigfoot didnt make sense. there wasnt that much media to store, computers had mostly a single HDD so a bigfoot was a pain the ass performance wise. but now, in the era of SSDs, and huge HD media, they would make sense. at least for that kind of use.

    • Elsoze
    • 5 years ago

    Does anyone know of any newer/updated RAID technologies that have come about to try to handle the increase of individual hard disk space? Even with 3/4TB drives people are justifiably squeamish about using RAID6 due to the slow read/write speeds of hard drives.

    I’d think that with all the processing power and caching available (via battery backed RAM and SSDs) that we’d have some kind of new RAID type or system to handle it. I suppose it may be more feasible to setup your own ZFS cluster loaded up with RAM than use a hardware based solution at the moment, but I don’t know enough about what’s current.

    • egon
    • 5 years ago

    Seeing as the Backblaze report has come up again, I’d like to query why the impressions it has left are so inconsistent with these figures:

    [url<]http://www.hardware.fr/articles/920-6/disques-durs.html[/url<] That's part of a long-running, ongoing series on hardware return rates (hardware.fr purports to have a major French retailer as a source), but it doesn't get much attention in English tech discussion due to the language barrier. Obviously, return rate isn't a direct indicator of failure rate (the Backblaze report was hardly an ideal comparison either). But some correlation would be expected, and if Seagate was such an outlier in terms of reliability, it seems to me the hardware.fr figures should reflect that much more than they do.

      • MadManOriginal
      • 5 years ago

      I haven’t looked at the Backblaze report in a while but iirc the Seagates in it were older drive generations that were known to be poor. I think they even said that they have too few newer Seagates to draw conclusions, and yet people like to dumb it down to brands because that’s how they present the data in graphs.

        • Waco
        • 5 years ago

        Their favorite drives these days are Seagate drives.

          • MadManOriginal
          • 5 years ago

          Yes, but they specifically say it’s because of price and despite their past data. That’s fair enough, it just means that people shouldn’t draw the conclusion of ‘all Seagates suck’ from the report.

            • Waco
            • 5 years ago

            The majority of their failures were from a firmware bug that hit after a few years of operation.

            From my experience so far with ~700 of the newer Seagate NAS drives…they’re just as reliable as any other large consumer drive.

      • DarkMikaru
      • 5 years ago

      No matter how you cut it, it comes down to personal experience. In my experience, Seagate & Hitachi drives have let me down far more than WD & Toshiba by a large margin. To the point where I just refuse to purchase anything Seagate anymore.

      In the end, that is all it comes down to. Personal experience / preference.

    • Sargent Duck
    • 5 years ago

    That’s a lot of porn…

      • UnfriendlyFire
      • 5 years ago

      Not if they’re 4K or 8K PNG photos.

        • chΒ΅ck
        • 5 years ago

        Or, you know, RAW…

    • Wirko
    • 5 years ago

    If any of the engineers who invented the original hard disk drive at IBM is still alive, they must feel really old now.

      • danielravennest
      • 5 years ago

      I’m not that old, but the first serious computer I worked on had a 10 MB IBM drive the size of a washing machine, with a 14 inch x 6 platter removable disk cartridge. I think the platters were double sided, so that makes 1385 square inches of disk surface, or 57 kilobits per square inch. That’s 240 bits per inch, about what smartphone screens do these days for pixel density.

      It’s truly amazing that capacity has gone up nearly a million times, while physical size has shrunk about a thousand-fold. Thus density is now around a billion times more.

        • tootercomputer
        • 5 years ago

        Add the factor of price per mg or whatever metric unit, and it’s even more amazing. I still have my 40mg Mitsubishi had from 1990, 200 US, typical price for that time. It’s relatively huge given its capacity. yeah, storage gains have really been amazing.

          • UberGerbil
          • 5 years ago

          $/MB were much higher just a few years before that: [url=https://c2.staticflickr.com/8/7251/7465276906_e22f29ee4c_z.jpg<]1986 PC HD prices[/url<] (10MB and 20MB)

    • Chrispy_
    • 5 years ago

    Yawn….
    Wake me up when they’re $[i<]x[/i<] a GB No, wait - that's not it. [i<]x[/i<]TB is a lot of data to lose in one go No, wait - that's not it either. Oh yeah, I remember: When is seagate going to release a "get perpendicular" style musical infomercial about [url=http://upload.wikimedia.org/wikipedia/commons/thumb/0/0f/Trigeminal_herpes_with_uveitis_and_keratitis.jpg/220px-Trigeminal_herpes_with_uveitis_and_keratitis.jpg<][i<]shingles[/i<][/url<]? (<-- not a link for the squeemish)

    • tootercomputer
    • 5 years ago

    Three thoughts here:

    1. It still amazes me to some extend what mechanical hard drives do, their tolerances, for an “old technology. Hats off to engineers.

    2. I used to purchase only WD, but I’ve had a few Seagates recently and have been pleasantly surprised.

    3. I don’t know about anyone else, but I have no pressing needs (nor do I expect to have any in the foreseeable future) for any additional storage drives. As I have upgraded to SSDs, and in some cases replaced HDDs with SSDs, I have a glut of empty HDDs some of which I’ve used for storage. They all run well, so no need for purchasing any new ones.

      • DarkMikaru
      • 5 years ago

      Personal experience is everything I guess. Over the past 6 years it seems every drive I have had to replace for family, friends, clients or myself have been Seagates. No lie. The final straw was a system I built for a neighbor about a year ago, drive lasted 2 months exactly before dying. Matter of fact, I’ve had several new Seagate drives die in that amount of time.

      Having said that though, I did recently experience one of my 2TB WD Greens fail after only a month of life. But that 1 WD failure over the last 6 yrs is nothing compared to the 10+ I’ve seen in that same time frame. Go fig.

        • tootercomputer
        • 5 years ago

        That’s amazing. My experience with WD has been good over the years, only one real failure. I just ignored Seagate, but I happened to get a couple of their drives for free, one an external drive, and both have been solid. But like you said, personal experience is everything.

          • nanoflower
          • 5 years ago

          The WD greens had a real problem due to their constantly spinning up and down that led to early mortality. I don’t know if that’s been fixed but hopefully so because it led to many of the bad reviews of the Greens.

        • Spunjji
        • 5 years ago

        I can mirror this experience. The only company I knew to fail more frequently was Maxtor. I’ve had occasional failure from all brands, of course, but Seagate are the absolute stand-out. Their early 320GB and 500GB 7200rpm notebook HDDs were even more “spectacular” in that regard.

    • ronch
    • 5 years ago

    It’s kinda ironic how the hard drive company that has the worst reputation in terms of product reliability usually (AFAIK) ships the largest capacity drives first.

    • Prestige Worldwide
    • 5 years ago

    That’s 8 TB I won’t buy

    [url<]https://techreport.com/r.x/2014_1_20_Backblaze_names_names_in_hard_drive_reliability_study/failure.jpg[/url<]

      • Takeshi7
      • 5 years ago

      If you extrapolate that graph, the Seagate 8TB must be really reliable because as capacity goes up, so does reliability.

        • Meadows
        • 5 years ago

        This also seems to imply that their existing 4 TB drives seem to be about as reliable as any WD drive.

        • Milo Burke
        • 5 years ago

        That makes as much sense as this [url=http://xkcd.com/605/<]XKCD comic[/url<].

      • Krogoth
      • 5 years ago

      You realize that Seagate ships far more HDDs than WD, Toshiba and Samsung? You would expect such figures assuming reliability is roughly equal on a per unit basis.

    • Farting Bob
    • 5 years ago

    Thats a lot of ones and zeroes. As always, i am glad when the higher capacities come along because it pushes the next tier down in price and i need a new 3 or 4TB drive soon.

    • Concupiscence
    • 5 years ago

    And it will fail just in time for the competitors to ride to the rescue.

    • flip-mode
    • 5 years ago

    Now we can lose our data 8TB at a time.

      • MadManOriginal
      • 5 years ago

      Just like people always say ‘back in the day, I paid $$$ for x’, I think people have been mentioning losing bigger chunks of data forever as well.

        • LocalCitizen
        • 5 years ago

        yup. i said that about DSHD

      • ClickClick5
      • 5 years ago

      Coming from Seagate, you betcha!

      • Ochadd
      • 5 years ago

      That’s so much data to be lost at a single point of failure… Could hardly guess what a RAID rebuild would take for 8TB of data in an active array. More than a week I assume?

        • sjl
        • 5 years ago

        Well, in my experience, you should allow 24 hours for a SATA drive (that’s how long the restriping took on my NAS when I added one or two 3TB drives to the RAID 6 array.) With these drives, a bit of back-of-the-envelope calculations, combined with sticking a wet finger up in the air, suggests that a week is not an unreasonable amount of time to allow, Just In Case.

        With that sort of time frame for a drive rebuild, I’d be looking seriously at three drive redundancy, rather than RAID 6 (never mind RAID 5) – ZFS FTW! Not cheap, but the pain factor in the event of array loss is too great to ignore, IMO.

      • trackerben
      • 5 years ago

      8^B

      • anotherengineer
      • 5 years ago

      pffff no way, 4 drive RAID 0, 32TB at a time or go home πŸ™‚

        • Farting Bob
        • 5 years ago

        RAMDisk during a thunderstorm is the only way to go.

          • UnfriendlyFire
          • 5 years ago

          And set it so it only saves the data to a 5400 RPM HDD when the computer shuts down.

          • UnfriendlyFire
          • 5 years ago

          Oh, and make sure to use a $15 Leadman PSU rated at 600W!

      • Wirko
      • 5 years ago

      Worry not, we also have technology available that lets us lose 1.44 MB at a time.

        • UnfriendlyFire
        • 5 years ago

        For a second I thought you were talking about the Zip drives.

      • Krogoth
      • 5 years ago

      Pffft, come back again when you lose a NAS box full of such drives. πŸ˜›

        • UnfriendlyFire
        • 5 years ago

        To CryptoLocker?

      • gmskking
      • 5 years ago

      Now we can lose our data 1TB at a time.
      Now we can lose our data 2TB at a time.
      Now we can lose our data 3TB at a time.
      Now we can lose our data 4TB at a time.
      Now we can lose our data 5TB at a time.
      Now we can lose our data 6TB at a time.
      Now we can lose our data 7TB at a time.
      Now we can lose our data 8TB at a time.

      • BIF
      • 5 years ago

      [quote<]"Now we can lose our data 8TB at a time."[/quote<] This is a silly comment that gets us nowhere in the long run. We should be demanding that our hardware and software makers keep striving for greater and greater capacity, especially as our total usage continues to grow. Mine does. 3D graphic libraries, virtual music instrument samples, and my own content such as photos, videos, animations, and art images; it all takes ever greater amounts of storage. Motherboards have a practical and finite number of SATA connections. Sooner or later, adding a new drive is not a practical solution. Replacing with a larger drive (or a RAID team of larger drives) is more practical, and is ultimately more economical and even uses less resources, if you really want to go there. I would like to have two or three 8TB drives for use as backups of my 2-4 TB operational systems. I find backup drives with capacity of 2X to 4X of my operational system usage actually work great, because they offer ample room for multiple full + incremental images. But that's just one of many reasons we should be excited about bigger drives.

      • DarkMikaru
      • 5 years ago

      Nope! That’s why you buy two! Problem Solved πŸ˜‰

    • UnfriendlyFire
    • 5 years ago

    Can’t wait for Backblaze to do some testing.

    On a side note, Backblaze did state that their favorite hard drives were Seagates, and they explained that even though Seagates had lower reliability, they were cheaper than competing hard drives and their RAID setups allowed them to quickly switch failed hard drives. They stated, “Buy the cheapest components and be able to quickly swap them.”

    According to their calculations, a hard drive that is twice as reliable is only worth an extra $0.16 per GB.

      • Waco
      • 5 years ago

      When you’re paying $.04 / GB for bulk storage an additional $.16 / GB is a MASSIVE jump in cost.

        • UnfriendlyFire
        • 5 years ago

        It was suppose to be $0.016. Forgot a zero.

        Here’s the link to the article: [url<]http://www.theregister.co.uk/2014/02/17/backblaze_how_not_to_evaluate_disk_reliability/[/url<]

          • Waco
          • 5 years ago

          Ha, that’s a bit different. πŸ™‚

      • Prestige Worldwide
      • 5 years ago

      Ehh, I have managed to snag 2TB Toshiba (Hitachi) drives for $79.99 CAD, more reliable drives than Seagate can be had for less than a Seagate drive if you wait for the right sale price to buy.

        • nanoflower
        • 5 years ago

        Microcenter has been selling those 2TB 7200RPM Toshiba drives for between 67-70$ US for months now. I’m half expecting to see a price drop before Christmas. Only problem is I’m not sure you can get them through mail order from Microcenter.

          • Prestige Worldwide
          • 5 years ago

          I’m in Canada, eh.

          79.99 is as low as they’ve gone in the great white north as far as PC Part Picker’s graphs will show me for the last 180 days. We aren’t blessed with the merciful Lord MicroCenter.

          [url<]http://ca.pcpartpicker.com/part/toshiba-internal-hard-drive-hdkpc09?history_days=180[/url<]

        • Waco
        • 5 years ago

        Seagate drives are just as reliable as any other. Aside from the bad firmware from a few years back they’re no worse than any other.

        If you really want deals on bulk drives check out bulk sales of refurb drives on Ebay. I snagged five 2 TB WD Purple drives for $50 each a few months back.

      • Liron
      • 5 years ago

      It’s only worth $0.016 more for that specific company because of what it costs them to replace a hard drive (given the salary they pay to a full time employee replacing hard drives all day.) If it takes longer than 15 minutes to replace the hard drives in your data center, or you have different salaries, then the figure does not apply. For my home computer, it takes over 72 hours to replace a hard drive, since I have to request the RMA and wait for the replacement drive to be shipped, so doubling the reliability would be more valuable.

        • Spunjji
        • 5 years ago

        Quite right. the considerations (and the consequences of data loss) are completely different for home suers and datacenters.

    • SoM
    • 5 years ago

    how much is this drive supposed to cost?

      • Pzenarch
      • 5 years ago

      The press release doesn’t say, but I wouldn’t worry about that just yet. While the lesser spotted 8TB drive remains niche, the old rule of pricing applies. If you have to ask, you can’t afford it (or don’t want to pay the 500% markup on it) πŸ˜›

Pin It on Pinterest

Share This