Sony's 185TB tape: What does it mean for storage?

Sony's 185TB tape: What does it mean for storage?

Summary: The R&D guys at Sony announced a real breakthrough: Nanotech that enables uniform crystalline orientation on tape - and ultra-high density. Will it save tape?

TOPICS: Storage, Hardware

This is cool technology: Sony researchers developed a soft magnetic underlayer using sputtering, a deposition technique commonly used for chips. Here's a Sony diagram of the tape cross-section:

Courtesy of Sony Corporation


The underlayer made it possible to deposit 7.7nm magnetic particles uniformly on the tape. The particle's small size and uniformity enabled the record-smashing 18.5GB/in<sup>2</sup> density.

The Storage Bits take
Demonstrating this density is a great start: It gives customers a reason to stay with tape for the next 10 years. But unless Sony can accelerate time to market, we're not likely to see it in a product.

New tape formulations also require new heads optimized for the nanotechnology. New heads are complex and expensive to develop.

The read/write electronics will also have to be able to handle 74 times more data if tape speed remains constant. Tape drives will have to achieve even higher precision required for much denser tracks — and precision mechanics are not easy or cheap.

They'll also have to answer for durability and robustness. Today's LTO tapes only support 200 complete reads or writes, making 3-bit per cell flash look like granite by comparison. And how will the tape hold up in long-term storage?

While the economics of the tape business are trending against it, this is a magnificent technical achievement. It shows us what is possible with this venerable media type, but not what is likely.

Comments welcome. Where would 185TB tapes make sense in your organization?

Topics: Storage, Hardware

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • The place I worked stored data in three tiers..

    1. on line and immediately available for processing
    2. nearline - on tape, but accessable in 10-20 seconds.
    3. archive - on long term tape, accessable anywhere from 10 -20 seconds (still in the library), 30 minutes (on a rack in the storage area), 3 days (offline external storage requiring a trip to get).

    The data files are also segregated - small (under 3 GB). medium (3-50 GB), and large (50 and up GB). The small files would go to the smaller tapes without a known read/write limit other than detected error rate. The medium would go to larger tapes, and the archived data would go the largest tapes for near permanent storage.

    The small/medium files would be transfered from tape to tape when the deleted/superseded data exceeded around 40% of the tape - thus recycling the tape through read/write error checks. The large files would also go through this stage, but when the data aged over a year it be moved to long term storage.

    Since this was an oceanographic site, the modeling worked on both weather and ocean currents, and had to keep archived data for (nearly forever) as models get updated, they would reprocess the original data, and compare the results to reality, and to previous models to verify improvements.

    Data sets in the 100GB range were rather common.

    When I changed jobs, they had reached around 400TB of archives, and increasing at about 1TB per month (and the rate was increasing as computers got faster, and the models got more detailed).

    I would imagine that the "cloud" service companies will need a LOT of storage, and disks/ssd just isn't cost effective for long term storage.

    Do I expect a single tape to hold 185 TB? no. I would expect that to be broken up into maybe 50TB units, making average access to data much faster - and increase the lifetime of each tape. Even where I was, tape was also used in a mirror form - two or more tapes would hold the same data, providing long term redundancy and recoverability (again, extending the lifetime of the tape as a byproduct).
  • It means Nothing -Really Sony is just Pissing $ away

    Sony is Constantly trying to fight the Betamax Wars- It has nothing to do with the Technology and everything to do with established Standards for Tape Backup. LTO Ate AIT and SUPERAIT's Lunch and killed off Proprietary Tape Formats LTO 7 and 8 are already being Talked about as IBM announced they are working on a 125TB solution with Fujifilm. Nobody wants to go back to Proprietary non backwards compatible formats.. As cool as it sounds, this baby is stillborn.
  • It gives customers a reason to stay with tape for the next 10 years.

    Who are you kidding? Oh, sure, I'll use 1990's technology (read "tape") to protect my data, after all, its no big deal, right?

    Perhaps for the huge "cloud" data centers, this may be economical but for the mainstream it's just a new twist on old technology.

    TB prices on hard drives are cheap. I know my data is safe when its backed up to a 16TB NAS drive. I know I can restore that data quickly (if not immediately via visualization).

    Tape? Yea, right. Tape stretches making it unreadable. Tape deteriorates easily making it unreadable. Tape is dependent on specific hardware making it unreadable if the drive fails. Tape takes HOURS at best to restore anything of significant size.

    Tape? Oh please. You're not still running Windows 3.1 by any chance, are you?
    • Tape is still way relevant.

      We still use tape and in fact have increased our tape footprint. It's the only economical way to store the hundreds of terabytes of images that we produce. These big movie studios shooting in HD and beyond have movies that go up in to the hundreds of terabytes. That's hundreds of terabytes for a single movie!. They might initially store it on a huge NAS system initially but it eventually gets put on tape. Tape is rated to last for many many more years than a HDD is. If you do the math on how often you have to replace a drive on an enterprise level NAS and how much electricity it takes to run it tape is still the undisputed winner. I'll concede that there is a point where Tape doesn't make sense. If you aren't storing a large amount of data then I would probably just do a disk to disk archive and call it a day but when you are doing 100TB + then you need tape!
  • Just what I need!

    I should be able to store ALL my VIC-20 BASIC programs on a single tape!
  • Sounds like something the Mormon Church's genealogy project could use.

    They add something like 15TB PER DAY. A single tape that could handle a WEEK'S worth of data?!? Good deal!
  • Trust tape?

    Over the years I've had almost continuous failures with many formats of tape. When I really needed to restore it was a nightmare. Since then, I image to a huge hard drive and every few weeks delete the old images. Never failed when I needed it.
  • high-end tape still useful

    Were I work we still use tape for backup, particularly to have off-site backup which isn't connected to electricity. And it's been very reliable for us, using enterprise-level tape storage.

    This new technology is pretty remarkable, but given the density I can't help but wonder if they'd build it into tapes with a smaller physical size. We have enough data to store on tape that I can't see us switching to disk for all of it, but that doesn't mean we have to have 185 terabytes per tape. Right now we're using LT06 tapes which hold 2.5-TB of uncompressed data (iirc), so something like 16-TB tapes would still be a big improvement for our purposes. Much over 32-TB per tape would probably be counter-productive for the way we use tape.

    Some useful information on current linear-tape technology:
  • Also, obviously for large operations

    To key off an earlier comment: if all you have is 16-TB of data, then you obviously aren't going to waste the money to purchase enterprise-level tape storage. But when you need multiple backups of data (onsite and offsite), and your data is *growing* by 10-20 terabytes per week, then tape still a cost-effective way to get the job done.