100,000 year nanowire storage

100,000 year nanowire storage

Summary: For all the technology we use to store data, there is one problem that has no good solution: longevity. Some scientists at the University of Pennsylvania - home of some of the first computers - have developed a new kind of memory that is 1,000x faster than flash and should hold its contents for 100,000 years.


For all the technology we use to store data, there is one problem that has no good solution: longevity. Some scientists at the University of Pennsylvania - home of some of the first computers - have developed a new kind of memory that is 1,000x faster than flash and should hold its contents for 100,000 years.

Phase change memory The Penn folks have come with a new wrinkle on phase change memory which is already used in recordable CDs and DVDs. The phase changes from a disordered amorphous state to a an ordered crystalline state. Resistance and reflectivity change as well and encode the 1's and 0's of binary data.

The goal is to build phase change memory chips instead of optical media. Intel, Hynix, Samsung and others have licensed the technology from Ovonyx, a company dedicated to commercializing the technology.

So what's cool? The new technology has several advantages:

  • The phase-change nanowires are self-assembling, so they are simpler than traditional semiconductor manufacturing, where Ovonyx is focused.
  • Data writing, erasing and retrieval time of 50 ns - 1,000 times faster than flash - fast enough to replace DRAM in many applications.
  • Scalable to terabit - flash is currently a few gigabits - chip densities
  • 100,000 year data retention - assuming anyone is around to read it then

The only not-cool thing: the research leader, Prof. Ritesh Agarwal, predicts it will be 8-10 years before commercial products ship. Darn and double-darn.

The Storage Bits take Flash is a great technology, but it's American to want something better. Feature sizes are getting so small that even deep ultra-violet semiconductor processing is coming to the end of the road. Nano-scale self-assembly, as at Penn and in my prior post - Engineering the 10 TB notebook drive - is promising continued progress without relying on electron beam lithography and other exotic techniques.

Storage progress will continue and I, for one, couldn't be happier. For more on the work at Penn read the article at Physorg.com.

Comments welcome, as always.

Topics: Hardware, Networking, Processors, Storage

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
  • More on faste nanowires

    Good morning,

    I already wrote about these faster nanowires last month on my ZDNet blog. Please check "Faster memories made of nanowires" (http://blogs.zdnet.com/emergingtech/?p=695) for more details.

    Roland Piquepaille.
    Roland Piquepaille
  • And this is why I think HD Optical DIsk Formats are DOA

    Sounds perfect!
  • RE: 100,000 year nanowire storage

    Petabytes to Infinity on rewritable optical
    holographics and spintronics having 5ns atomic switch times is already in testbed and will be
    out in 3 to 5 years.


    I will be amazed at how they can handle the power, heat, cross talk, and impedance problems just to mention a few.

    Also most phase change material is ferroelectric and it will be interesting to see what patent conflicts will arise by the FERAM and FRAM folks.

    The claim of 100,000 year shelf life is funny.
    • he claim of 100,000 year shelf life is funny....

      I agree I have floppies that are still good, but no floppy drive to put them in. I setup a system and just copied them all to CD for archiving(I had to borrow the floppy drive). So they think the hardware will to read this storage will last 100,000 years ????
      • The claim of 100,000 year shelf life is funny...

        Actually, I'd settle for 20 years. That would keep all the stuff I'm required to keep by law. The only problem, as has been mentioned, would be to have a) the ability to read it, and b) a running program which could use it. Kind of like the music device from the Krell in the old SiFi movie "Forbidden Planet".
        • Forbidden Planet...

          I liked that movie.
  • RE: 100,000 year nanowire storage

    Let's forget jokes about theoretical 100K yr. storage and focus on more practical periods. Certainly there are archives that might benefit from storage cycles of 100 - 1000 years. One should not have too much trouble conceiving of devices for reading these media that could be preserved along with the storage media itself. In fact, these devices and the storage might be economically integrated into a single package. Just because innovations sound outlandish, they should not inspire derision. Or worse: remember the Inquisition?
    • Right now the technology sounds good...

      But as we go forward Nanowire storage will fall victim to the next great thing and there will be no systems capable of reading it.
      Look at all of the old MFM RLL drives, even if you kept the controller card it would be hard to find an ISA bus computer to plug the controller into.
  • Have we gone this way before?

    Some of the historical puzzles we've come across in archaeoligical digs suggest that 1000 years from now (never mind 100,000 years) anyone digging up well preserved CD and DVD discs would speculate on their mirror qualities and be unable to figure out anything that they have because the readers would have turned to dust centuries before. Even events carved in stone fail to inform present day archaeologists what they're all about
  • RE: 100,000 year nanowire storage

    This could create a new paradigm shift in system and storage thinking. With say, a few Tera bytes of non-volatile memory, would your storage subsystem just be a part of shared system memory? Or, would it be the other way around in that your system memory could just be shared memory from your storage system? And then what about video memory?

    Stuff to think about.
  • Porn is now forver!

    Long live the massive porn collection!
  • RE: 100,000 year nanowire storage

    Nanowires, bubble memory, fluid memory, etc etc..

    We can't even read data that was put on video tape from the early 20th century because there are no machines to read the width of tape that it was created on nor can we read the data that was put on the very large flexi disks.. remember them?

    Nanowires... just put the darn data on a mountain of paper (yes that medium has been around for thousands of years already).