Fujistu To Cram 1PB of Storage Into a Single Tape Drive

News
By
published

Fujifilm expects LTO capacities to grow at rapid pace for at least 15 years.

IBM
(Image credit: IBM)

An average tape drive stores roughly 12TB of data, but Fujitsu is working to expand that to 1PB, or 2.5PB after compression (2,500 TB), by 2035. That's a massive increase from the 30TB of storage on today's average tape drive. 

Most folks aren't aware, but the lion's share of the world's data is still stored on tape cartridges buried deep in data centers, and that trend isn't expected to change any time soon. The 8th generation of the linear tape open (LTO) open tape format is widely used in data centers today. Specced for over 30TB of compressed storage and up to 750 MBps data transfer rates, LTO-8 tapes allow datacenters to store zetabytes of data that will be available for decades to come. But as demand for storage is growing, tape developers — just like developers of other types of storage — need to adopt new technologies to increase the capacity of their products. 

Fujifilm this week presented its long-term roadmap for tapes that spans 15 years and includes 1TB cartridges featuring new materials, as well as a new magnetic recording method.

Today, tape cartridge makers use tapes with a magnetic layer consisting of Barium Ferrite (BaFe) magnetic particles. Such cartridges can store 12TB native or 30TB of compressed data, and their capacity is expected to gradually increase to about 100TB by 2026. In November, Fujifilm and IBM unveiled a new tape technology featuring a Strontium Ferrite (SrFe) magnetic layer that promises to enable future cartridges to store up to 580TB of data by around 2029. 

The tape evolution naturally does not stop there though. Fujifilm now says that it's on track to offer 1 petabyte (PB) cartridges by approximately 2035. 

(Image credit: Fujifilm via ComputerBase)

In a bid to increase tape storage capacity further, Fujifilm plans to use tapes featuring a magnetic layer based on epsilon iron oxide nanoparticles (also known as Epsilon Ferrite) that are very small and can enable extremely narrow tracks as well as a very linear density, ComputerBase reported today. To properly record data into very small pitches, the magnetic properties of the material have to be altered. To switch the magnetic pole direction of Epsilon Ferrite, Fujifilm plans to use a method called focused‐millimeter‐wave‐assisted magnetic recording (F‐MIMR) that uses terahertz (THz) light. 

In addition to increasing recording density and tape capacity, F-MIMR also greatly increases writing speeds. Typical magnetic recording time on tape is on the nanosecond (ns) scale. Meanwhile, F‐MIMR is expected to write on the picosecond (ps) scale, according to Advanced Materials.

(Image credit: IBM)

Both Epsilon Ferrite and F-MIMR are fairly new technologies, so Fujifilm does not expect to use them commercially for about 15 years. But once the technologies are ready for primetime, Fujifilm expects to be able to build 1PB cartridges, which, considering a 2.5:1 ratio, will be able to store around 2.5PB of data, as noted by Blocks & Files. Even at 1PB, such tape cartridges would significantly exceed capacities offered by today's hard drives. 

See more HDDs News
Anton Shilov
Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

6 Comments Comment from the forums
  • hotaru251
    how long would it even take to recover if it died....thats a massive amount of data to risk on 1 drive.
    Reply
  • spongiemaster
    Both Epsilon Ferrite and F-MIMR are fairly new technologies, so Fujifilm does not expect to use them commercially for about 15 years.

    15 years? It took us less than half that time to develop the atomic bomb and land on the moon. Besides Half-Life 3, how many things are in development for 15 years?
    Reply
  • cryoburner
    spongiemaster said:
    15 years? It took us less than half that time to develop the atomic bomb and land on the moon. Besides Half-Life 3, how many things are in development for 15 years?
    What would be their motivation for rushing development? An accelerated development process would likely cost a lot more to achieve a similar result, and would just move them closer to the point where improvements in density slow to a halt. Unlike those kinds of government programs, a company like this needs the end result to be profitable.

    Though I must say, the article is a bit nonsensical, reporting on a tech company's plans for 15 years down the line. These kinds of articles tends to not age all that well as things often get delayed or modified. In 15 years, how many of the company's employees will even be working there still? Fujifilm might even be merged into some other company or sell off their tape storage division by then. Or maybe some alternate technology will be developed that makes tape storage obsolete by that point.
    Reply
  • GenericUser
    spongiemaster said:
    Besides Half-Life 3, how many things are in development for 15 years?

    I was going to say Duke Nukem Forever, but that was only 14 years. And for what we got for it...

    As for the HL3, I think Valve gave up on that a long time ago, for the variety of speculated reasons floating around for it. A man can dream though.
    Reply
  • spongiemaster
    cryoburner said:
    What would be their motivation for rushing development? An accelerated development process would likely cost a lot more to achieve a similar result, and would just move them closer to the point where improvements in density slow to a halt. Unlike those kinds of government programs, a company like this needs the end result to be profitable.
    And you think it would be easier to get a return on development costs by dragging it out 15 years? Who has a roadmap that long? Think about how many companies have dropped out of the storage market in the last 15 years.
    Reply
  • cryoburner
    spongiemaster said:
    And you think it would be easier to get a return on development costs by dragging it out 15 years? Who has a roadmap that long? Think about how many companies have dropped out of the storage market in the last 15 years.
    A lot of this is going to be down to gradual refinements of the process over time. Sure, it might be possible for them to move quicker, but again, the total development costs would likely be higher, at least in the short term, and tape storage is ultimately all about providing storage at the lowest possible cost. As long as they manage to keep up with, or stay slightly ahead of the competition, they don't have much incentive to rapidly push development forward. Companies are going to buy their product either way. And if they did rush development, only to find that some alternate technology made what they developed obsolete, they wouldn't have the option of shifting their course of development down the line to something that would be more practical, potentially resulting in money wasted on developing something they won't be using.
    Reply