IBM/Fujifilm prototype tape hits 154TB capacity

IBM/Fujifilm prototype tape hits 154TB capacity

Summary: Just weeks ago, Sony announced a lab demo of a 185TB tape. This demo uses technology similar to today's tape, but is more refined. Will tape win back some ground lost to disks?

TOPICS: Storage, Hardware, IBM

Magnetic tape is less than a century old and, for a while in the 1950s, it was the main mass storage for computers. Hulking tape drives dominated early computer rooms — especially in Hollywood movies — and were finicky combinations of electronic, mechanical, and pneumatic parts.

By the early 60s, faster and only slightly less finicky disk drives replaced tape as primary storage. Despite many predictions of tape's demise, it's still here.

Fujifilm and IBM

The underlying technical improvements behind this demo are interesting:

  • New high-density, dual-coated particulate magnetic tape. A new ultra-fine, perpendicularly oriented barium-ferrite magnetic medium without costly sputtering or evaporation coating
  • Enhanced write field head technology. IBM Research-Zurich, with help from IBM San Jose, developed a new write head with much stronger magnetic fields that works with smaller volume magnetic particles adding increased coercivity necessary for long-term stability.
  • Advanced nano-scale head positioning. The 27-fold track increase over the LTO6 format requires a new low noise tape transport system, an improved servo pattern and method for detecting position information at nano-scale resolution, and a new track follow controller. They achieved a track-follow performance with a standard deviation of only 10.3 nanometers.
  • A new signal processing data channel that enables the reliable detection of data at a linear density of 600,000 bits per inch with a 90nm wide GMR read head. All this AND the same user bit error rate as the latest IBM enterprise tape drives

10.3nm? That's smaller than today's chip feature sizes! Here's how the tape roadmap compares to other media:

Screen Shot 2014-05-19 at 9.53.22 PM
Graph courtesy IBM


The Storage Bits take

The IBM/Fujifilm team thinks they can keep doubling tape density every two years for the next decade. That's faster density growth than disks will achieve, which will nudge the economic advantage back towards tape.

Tape's Achilles heel is, of course, access time. It works for rarely accessed data — archives — but its glory days of the 1950s, where it was primary storage, will never return.

But this demo — and Sony's — shows there is still life in the old technology yet. Predictions of tape's demise are, once again, premature.

Comments welcome. Do you use tape? Please tell us how and why.

See also:

Topics: Storage, Hardware, IBM

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 reason I use tape still

    I don't know of any other modern storage medium that is suitable for offline. Please someone tell me if I am wrong because I would love to know.

    If you take an hdd offline and stick it on a shelf chances are it won't spin up in less than a year. Sure the data is still on the platters but bearings seize and hdd's were just never designed for offline applications. Ssd's are just too expensive and their persistence when offline is questionable. Optical media is a minefield for the ignorant or uninformed.

    Tape is the only medium I can stick on a shelf comfortably knowing that in 5, 10, maybe even 20 years the tape and data will still be readable.
    • More likely than not though, in as few as five years, ...

      ... and certainly in twenty years, there will be no tape drives which will be able to read that data. You are certainly correct about HDD but the same problem exists no matter what the medium.

      The Second Law of thermodynamics tells us that entropy always increases. Nothing has an indefinite shelf-life. CDs were once thought to have a practical shelf life of 20 years - and I have CD media older than that that still works.

      Modern removable media drives (BD+/-RW) will still read and write that 20-year-old CD media but now USB Flash Drives are replacing the highest-capacity Blu-Ray drives in new systems - in which USB and MicroSD are the only removable media options - with capacities far exceeding BD media.

      The only way to be sure that your archival media will be there when you need it is to periodically move it to newer technology. If you are going to do that, it seems to me that keeping multiple copies on-line (and off-site) is only marginally more expensive keeping if off line on a medium which remains readable as the technology advances.
      M Wagner
      • IOW

        U don't know any alternatives to tape for offline storage.
  • With today's multi-TB consumer-grade HDD technology ...

    ... with 3.5", 1TB+ HDD costing less that $0.10 per GB, I find it hard to believe that tape will ever come back in a big way.

    While tape may be more efficient for long-term archival storage, the availability of retrieval systems is very short-lived. In the last 30-some years, several generations of tape and disk archival storage technologies have come and gone and left behind mountains of media which cannot be read because NO ONE makes a drive which can read the data off of the storage media.

    In short, you have to keep moving your archival data to newer and newer technologies to keep it accessible so, arguably, you are better off maintaining multiple copies of the most valuable data in multiple on-line locations using the latest technology available. Even this will not insure that the data will be accessible forever - whatever that means - but does it have to be? Most of the data stored today is of no intrinsic value past the life of the person who stored it in the first place.
    M Wagner
    • Availability and reliability are two separate things

      But I digress. I honestly have no idea what you are talking about. Your viewpoint does not negate the need for offline storage.
  • Robin, do I understand correctly?

    I get the impression from the two articles that although Sony has slightly higher density, they're not anywhere near a marketable product, but IBM/Fujifilm has a prototype that could be turned into an end product within about two years.
    • No, this is more a proof-of-concept

      This density is the level they hope/plan to reach in 10 years - assuming the economics supports the development expense. That is the larger question to me, because object stores make it easy to upgrade in place, reducing the cost of keeping data very near-line.

      But if tape density increases faster than disk density for the next decade, then the tape guys have a chance to win back some market share. And disks are in a squeeze of their own from flash and future NVRAM technologies.

      R Harris
  • Tape is an insurance policy

    Tape is mainly used for backing up data as an insurance against disasterous loss of the primary storage. For this reason, once written the tapes are often stored offsite in fire safes, hopefully never to be seen again. That's why seemingly fragile tape is used; the cost per megabyte just can't be beaten. That's why tape still has a place in the storage hierarchy. Like all insurance you never really want to have to use it, but if you do then you need every single data bit to be valid. To be confident of this the read error rates must be negligably small, even taking into account interchange (the playback drive may well be a different unit from the one that wrote the tape), and the inevitable slow degradation in Signal to Noise ratio that will occur. You can't ever completely escape from randomly flipping magnetic domains by the second law of thermidynamics. For the required high level of confidence in the reliability of the storage the majority of the information stored on magnetic tape is certainly likely to outlast the availability of the hardware that can read it back. A great deal of effort goes in to designing for this longevity, interchange and reliability. I've heard stories of NASA having reels of 1/4" tape from the 60s on which every bit is still correct (after error correction) - the problem is keeping the reading hardware working.