Scientific discovery moves racetrack memory closer to reality

Dec 23, 2010

IBM researchers today revealed a previously unknown aspect of key physics inside Racetrack memory -- a new technology design which stands to improve memory capabilities within mobile phones, laptop computers and business-class servers. This new class of memory could enable devices to store much more information - as much as a factor of 100 times greater - while using much less energy than today's designs.

The Racetrack project -- which started in IBM's Research labs only six years ago -- flips the current memory paradigm on its head. Instead of making computers seek out the data it needs – as is the case in traditional computing systems – IBM’s Racetrack memory automatically moves data to where it can be used, sliding magnetic bits back and forth along nanowire "racetracks." This technique would allow electronic manufacturers to design a portable device capable of storing all the movies produced worldwide in a given year with room to spare.

Digital data is typically stored in magnetic hard disk drives, which are low-cost but slow due to their moving parts, or in solid state memory such as Flash memory, which are faster but more expensive. Racetrack memory aims to combine the best attributes of these two types of devices by storing data as magnetic regions – also called domains – in racetracks just a few tens of nanometers wide.

The new understanding, revealed today in the journal Science, allows the precise control of the placement of these domains, which the IBM team has proven can act as nano-sized data keepers that can not only store at least 100 times more memory than today's techniques, but can be accessed at much greater speeds. By controlling electrical pulses in the device, the scientists can move these domain walls at speeds of hundreds of miles per hour and then stop them precisely at the position needed -- allowing massive amounts of stored information to be accessed in less than a billionth of a second.

The full scientific details of the discovery can be read in the scientific paper. In short, the IBM scientists were the first to measure the time and distance of domain wall acceleration and deceleration in response to electric current pulses, which is how digital information is moved and processed in Racetrack memory. This not only gives scientists an unprecedented understanding and control over the magnetic movements inside these devices but also advances IBM’s Racetrack memory -- driving it closer to marketplace viability.

“We discovered that domain walls don't hit peak acceleration as soon as the current is turned on, and that it takes them exactly the same time and distance to hit peak acceleration as it does to decelerate and eventually come to a stop,” said Dr. Stuart Parkin, an IBM Fellow at IBM Research – Almaden. “This was previously undiscovered in part because it was not clear whether the domain walls actually had mass, and how the effects of acceleration and deceleration could exactly compensate one another. Now we know domain walls can be positioned precisely along the racetracks simply by varying the length of the current pulses even though the walls have mass”.

To achieve the densest and fastest possible memory, the domain walls inside the device must be moved at speeds of hundreds of miles per hour to atomically precise positions along the tracks. These timescales (tens of nanoseconds) and distances (micrometers) are surprisingly long, especially since previous experiments had shown no evidence for acceleration and deceleration for domain walls driven along smooth racetracks with current.

A closer look at Racetrack

For nearly fifty years, scientists have explored the possibility of storing information in magnetic domain walls, which are the boundaries between magnetic regions or "domains" in magnetic materials. Until now, manipulating domain walls was expensive, complex and used significant power to generate the fields necessary to do so. In a proof of concept paper in 2008(i) IBM researchers were the first to demonstrate the potential of , showing how the use of spin momentum considerably simplifies the memory device.

The details and results of this research effort will be reported in the December 24, 2010 issue of Science. The paper is titled, “Dynamics of magnetic under their own inertia,” and is authored by Luc Thomas, Rai Moriya, Charles Rettner and Stuart Parkin of IBM Research – Almaden.

Explore further: New multiscale model unifies physical laws of water flow to span all scales

More information: (i) Hayashi, M., Thomas, L., Moriya, R., Rettner, C. & Parkin, S. S. P. Current-Controlled Magnetic Domain-Wall Nanowire Shift Register. Science 320, 209-211 (2008)

For more information about Racetrack Memory, please visit: www.almaden.ibm.com/spinaps/research/sd/?racetrack

Related Stories

Beating the back-up blues

Apr 03, 2009

That sinking feeling when your hard disk starts screeching and you haven't backed up your holiday photos is a step closer to becoming a thing of the past thanks to research into a new kind of computer memory.

Racetrack memory

Nov 15, 2010

Imagine a computer equipped with shock-proof memory that's 100,000 times faster and consumes less power than current hard disks. EPFL Professor Mathias Klaui is working on a new kind of "racetrack" memory, ...

Walls falling faster for solid-state memory

Jun 09, 2010

After running a series of complex computer simulations, researchers have found that flaws in the structure of magnetic nanoscale wires play an important role in determining the operating speed of novel devices using such ...

Greener memory from random motion

Mar 01, 2010

Random thermal fluctuations in magnetic memory can be harnessed to reduce the energy required to store information, according to an experiment reported in the current issue of Physical Review Letters. The de ...

Recommended for you

Atom probe assisted dating of oldest piece of earth

1 hour ago

(Phys.org) —It's a scientific axiom: big claims require extra-solid evidence. So there were skeptics in 2001 when University of Wisconsin-Madison geoscience professor John Valley dated an ancient crystal ...

Could 'Jedi Putter' be the force golfers need?

Apr 18, 2014

Putting is arguably the most important skill in golf; in fact, it's been described as a game within a game. Now a team of Rice engineering students has devised a training putter that offers golfers audio, ...

Better thermal-imaging lens from waste sulfur

Apr 17, 2014

Sulfur left over from refining fossil fuels can be transformed into cheap, lightweight, plastic lenses for infrared devices, including night-vision goggles, a University of Arizona-led international team ...

User comments : 53

Adjust slider to filter visible comments by rank

Display comments: newest first

Quantum_Conundrum
3.7 / 5 (3) Dec 23, 2010
What the heck?

We are suddenly getting a tsunami of quantum, optical, and spin-based discoveries and even devices in the past week or so.

Is everyone suddenly paranoid someone else is going to beat them to their patents, so now they have to publish even if they were stalling before?

These discoveries are amazing. They are now saying they think they can store like almost a terrabyte in a Smart Phone's memory? And what? A petabyte in a device the size of PC's hard drive?

Moore's law was suppposed to be like 2^(N/2), but they just jumped two orders of magnituded within 6 years...
Husky
5 / 5 (3) Dec 23, 2010
i want to have 4 of those tracks in raid configuration put on the die of my cpu so that i could do away with the Level2 and 3 cache, the HD, the SD and the RAM all together, also i could see applications in videocards where you could have many parallel tracks to do stream processing, ultrafast shading etc
fmfbrestel
2.1 / 5 (9) Dec 23, 2010
Well, first off, there wont be petabyte drives in PC's because there is no need for that much space. (at least until program designers figure out how to use it) but it will be huge for portable computers.

IBM is also predicting metal air batteries that will last 10 times longer then current generation. With a week of battery time and a terrabyte of memory in phone (why do even still call them phones? actually making calls is the least of their ability) you wont even need a desktop computer, just keyboard maybe and a monitor.
Quantum_Conundrum
4.5 / 5 (4) Dec 23, 2010
If I'm understanding this uses nano-wires in the same manner magnetic tape has previously been used, except that the magnetic information is being moved along the device, instead of actually moving any parts. This allows the magnetic domains to move at incredible speeds limited only by the current and duration.

So why not also use this same technique to push magnetic domains, and therefore spin-bits, across nano-wire buses inside and between chips?

Oh yeah, look at the other IBM post.

Holographic video conferencing in five years. They'll be using those petabyte hard drives alright.
dnatwork
5 / 5 (10) Dec 23, 2010
No need for that much space? They said that back when the standard was 128k of RAM.

With that much space, someone will simply come up with a way to use it. Here's one: Real-life Tivo. Record everything that's going on around you 24/7, including your physical space but also all the data streams you encounter (phone calls, text messages, e-mails, TV shows, internet radio, facebook/twitter whatever-they're-called). Don't erase anything, just keep it around in case you want to rewind.

Why would anyone want this? I don't know, but I can't understand facebook, either.
Quantum_Conundrum
5 / 5 (2) Dec 23, 2010
Yeah, for example, holographic video of your family reunion.

...and record the Holographic video conference you had with the boss (or the employee.)

Starcraft 3 and Diablo 4 with holographic actors instead of dinky animated 3d models.
dirk_bruere
5 / 5 (4) Dec 23, 2010
Petabyte drive...
I really would like to have the best 10,000 movies of all time in *uncompressed* 1080p. That's around 40PB if my calcs are correct
Quantum_Conundrum
5 / 5 (2) Dec 23, 2010
Google and other data centers are going to be able to store...everything...about everything...and everyone...

Figure the "enterprise" versions of these devices are probably going to be like 100 petabyte or even exabyte storage devices specifically designed for use with servers.

A buddy of mine works for one of Google's competitors, and that company has converted an entire shopping mall to a data center, with row after row after row of servers throughout most of the complex. Imagine if all of them had these...

It's like the forbidden planet.
rgwalther
not rated yet Dec 23, 2010
The future is yesterday, but it is not forbidden.
fmfbrestel
not rated yet Dec 23, 2010
Yeah, i guess holograhpic displays would require huge memory capacity, but i just dont see that happening super quick. But if anyone would be in position to disagree it would be IBM, so what i do know?

Yeah, people might not like the cloud, but its coming. Google and Amazon and Apple are all building server farms as fast as they can right now. And they would all tell you that they still need more. It's a good time to be selling the servers and making the silicon inside them.
MorituriMax
3 / 5 (2) Dec 23, 2010
fmfbrestel, "Well, first off, there wont be petabyte drives in PC's because there is no need for that much space."

Hmmm, I remember someone saying once, "We'll never need more than 640K of RAM." Bill Gates if I recall correctly.
Yogaman
4 / 5 (1) Dec 23, 2010
Sure it's back-of-the-envelope, but petabytes in nanoseconds sounds about right for a nice, comfy Matrix pod. ;)

But is Racetrack Memory the winning technology? It's really, really hard to push the king off the mountain.
As ComputerWorld explains, http://www.comput...myId=147
commercial Racetrack Memory products are 2-5 years away.

In comparison, Flash has been doubling capacity about every 1.3 years while halving price about every year. (Does anybody remember how great bubble memory was going to be? Or Nanochip? Or how optical drives would supplant magnetic hdds?) For decades, clever disc engineers have pushed the bar higher.

Luckily, we non-combatant consumers win no matter what technology ends up on top. And now it looks like my next laptop will have solid-state mass storage, not rotating. Cool.

But flash or racetrack in 2-5 years? Depends on how clever the Flash engineers are, too. No?
Nemo
5 / 5 (3) Dec 23, 2010
I'm still in awe that I was able to buy a 2 TB drive last week for $85US. We live in amazing times.
lexington
5 / 5 (1) Dec 23, 2010
I'm still in awe that I was able to buy a 2 TB drive last week for $85US. We live in amazing times.


Yeah, it's pretty awesome. I can carry thousands of time more information in my pocket than could be fit in a whole building when I was born.
CSharpner
5 / 5 (1) Dec 23, 2010
fmfbrestel,

Well, first off, there wont be petabyte drives in PC's because there is no need for that much space. (at least until program designers figure out how to use it) but it will be huge for portable computers.

Are you KIDDING??? I can use any amount you make available to me. I can use up a petabyte in 612 days. If I hook up 5 HD security cameras around my house (one for each external wall (4) and one for the front door (5 total)), each records 13.6GB/hour, 326.4GB/day, 1.632TB per day for all 5. That'd take 612 days to use up a PT. Not to mention all the possibilities that open up with that amount of storage. I'd also have cameras with multiple views on my vehicles while driving... I'd probably have a camera on my body somewhere, recording my entire life, have screen recorders that record full video of everything I do on the computer, making it easier to use my computer screen like a DVR (rewind what I had done a few minutes earlier).

continued...
CSharpner
5 / 5 (1) Dec 23, 2010
continued...

Then imagine all those cameras recording in 3D. I don't know about the 3D codecs yet, but I'd imagine they consume maybe 1.2 to 2x as much space per unit of time.

Imagine your DVR... I'd have mine recording continuously. Today, that'd be 2 receivers continuously recording. I think my Dish Network DVR records about 2.6GB/hour in HD, so I could record 5.2GB/hour all day long, non-stop. That's 124.8GB/day, 45TB, year... just on 2 channels of continuous recording. It also opens up the doors to "black box" type data recording of driving history, temperature, humidity, wind, power usage, etc... The possibilities are limitless.

Trust me, when that type of storage is available, we programmers will use it up. I've got 4.5TB of hard drive space on my primary home computer right now, and I'm about out of space.
CSharpner
not rated yet Dec 23, 2010
Hmmm, I remember someone saying once, "We'll never need more than 640K of RAM." Bill Gates if I recall correctly.

That was my first thought too when I read him saying there won't be PB drives.

BTW, Bill Gates didn't actually say that, but that phrase has been incorrectly attributed to him for decades now.
CSharpner
not rated yet Dec 23, 2010
I'm still in awe that I was able to buy a 2 TB drive last week for $85US. We live in amazing times.

That's even better than the 1.5TB drive I got at Walmart 2 weeks ago for $70. Jeesh! We can now buy the storage equivilent of the human brain (~4TB) for $170. Freaking amazing! And that's just off-the-shelf tech in stores TODAY...

(yes, I know... we don't really know the storage capacity of the human brain. I quote 4TB because that's what I've heard from several sources over the last 10 years... just speculation, I know).
DamienS
5 / 5 (2) Dec 23, 2010
We can now buy the storage equivilent of the human brain (~4TB) for $170. Freaking amazing! And that's just off-the-shelf tech in stores TODAY...

(yes, I know... we don't really know the storage capacity of the human brain. I quote 4TB because that's what I've heard from several sources over the last 10 years... just speculation, I know).

Yes, that is very questionable, but the biggest difference between the human brain's memory and disk storage is that the brain's memory is highly associative, which easily trumps brute capacity alone.
EvgenijM
1 / 5 (1) Dec 24, 2010
No need for that much space? They said that back when the standard was 128k of RAM.

With that much space, someone will simply come up with a way to use it.


There is a limit to how much memory you need for a given task. Even tasks, such as video recording have an upper limit. For example - it is meaningless to record a movie at resolution which is higher than "resolution" of human eye. Eventually we will have more memory then we need to use for our task and goals. Heck even now I ask myself what to do with 350 gb of my 1 tb hdd.
Husky
5 / 5 (1) Dec 24, 2010
agreed, but upper limit recording and data sets with resolutions/sizes surpassing the human eye and brains will be usefull for computers to analyze astronomical data from telescopes/data from the LHC etc, history has shown that while solutions are mostly born out of acute problems/demand, the market will find or create a problem if they have to sell a solution in search of a problem
Husky
5 / 5 (2) Dec 24, 2010
i remember some guys working on the boomerang project a long time ago when RAM and HD was relatively expensive, they used 2 tokenring network cards in the same computer and 20 mtr cable as the racetrack to echo back data and virtially store data in the cableloop, it was slow sequantial access, it couldn't match with todays SSD speeds but it worked, i figure that with todays cheap Gigabit network cards a decent racetrack could be build by creative hackers
Quantum_Conundrum
5 / 5 (1) Dec 24, 2010
For example - it is meaningless to record a movie at resolution which is higher than "resolution" of human eye.


No it's not. For forensic purposes, you'd want a holgram with maximum resolution and maximum color depth.

Eventually we will have more memory then we need to use for our task and goals. Heck even now I ask myself what to do with 350 gb of my 1 tb hdd.


So you wouldn't set up perfect surveillance of your home if you had the means? That would use much more than 1tb, even with the best compression imaginable. Even if discarded anything over a month old, I'm pretty sure you'd need much more than 1 tb.
----

As for holographic codecs, it would depend on the camera, color depth, resolution, and optical array/lidar resolution.

Formula for surface of a square is X^2.

Formula for surface of a cube is 6X^2, while sphere is 4pir^2.

Because every pixel is mapped in 3d, you need (x,y,z) coordinate for every pixel in a hologram.

cont..
Quantum_Conundrum
3 / 5 (1) Dec 24, 2010
And so right away just in the mapping of holographic pixels, the storage space of each pixel's position is 50% larger.

Then to have resolution similar to a monitor, instead of 1600*900=1.44 million pixels, you would have like 1600*900*900 = 1.29 billion pixels.

You need all 1.29 billion pixels, even though only some are used in any given image, because you need to project the illusion of surfaces of unknown shape, curvature, and size in space.

So a hologram has thousands of times more pixels, and each pixel requires at least 2 more bytes of storage space just to record it's coordinate, which means a hologram file would be about 1350 times larger than a 2d photo, if it has 900 pixels resolution into the 3rd dimension.

Yeah, you'll be needing that petabyte hard drive.
irbis
1 / 5 (3) Dec 24, 2010
All the "laws of nature" open today, sooner or later be corrected by scientists. This is the same as the law of nature. For example unshakable before the SI system to date is disputed and corrected a number of scientists (an interesting example - http://physicalsy...ngl.html ). With regard to this article, really interesting study. We will wait for further results.
JimB135
not rated yet Dec 24, 2010
What a cool ride computer progress has been. My first hard drive was 20mb and sat under my then Mac Plus. About the size of one of those huge desk dictionaries it was loud and generated lots of heat. I thought it was the coolest thing ever. And it was at the time.

Upwards and onwards. Always waiting for the next cool thing.
Quantum_Conundrum
not rated yet Dec 24, 2010
As for data transfer, somebody somewhere is going to invent an optical port technology that works something like the laser scanner at the supermarket.

Your hand held gadget will still have wireless, but you'll also have "wired" connections which will use an optical port. You'll be able to place your phone in front of the port and transfer data at huge speeds over short distances, using arrays of optical transmitters and recievers based on the very same phased array concept currently used in radars, but will instead be with optics and infrared.

So the data transfer and the holographic cameras, etc, will all work based on similar principles, which are themselves based on phased array radar technology. The key difference is individual diodes are going to be on the nano-scale and micrometer scales where the wavelengths of visible light and infrared light operate
rgwalther
not rated yet Dec 24, 2010
The evolution of storage stops here! Byte me.
Merry Christmas
Quantum_Conundrum
5 / 5 (2) Dec 24, 2010
The evolution of storage stops here! Byte me.
Merry Christmas


I can think of "theoretical" ways to store more data in the same space. For example, an analog approach might allow you to stor data in higher bases, which translates to much higher data density.

DNA = 4 base
Codon = base 64

So if you could get to base 4 through some novel means of analog data storage, such as detecting and controlling the SIZE of the magnetic data points, and not just their existence (yes/no,) then you could store data in analog sets, allowing higher bases. Every time you double your base you double data density.

I'm not necessarily talking about DNA based computers, nor "analog computers". Just saying, you could have an analog converter for the secondary storage device.

Or in a photonic device, you could designate certain wavelengths of light to represent different values, 0, 1, 2, 3, 4, 5, 6, 7, etc, and then store data in "base n" in analog form at the nano-scale...somehow...
dtxx
5 / 5 (2) Dec 24, 2010
First of all, holograms are amenable to compression, and it can have an added measure of compression as well beyond 2d. For example, imagine a hologram of a 3d opaque sphere that fills 99.9% of the available pixels. You only need to represent the outer opaque layer and don't need to draw any pixels inside ever. Plus in this case the shape is easily represented as a vector.

So yes, the addition of a third coordinate as well as 900 pixels of z does add up to 1350x the amount of data. But your statement is untrue except in the case of a truly unusual image in which all pixels in the hologram are visible, random, and unique. In terms of real world and practical applications, you will not be storing images 1350x the size of an equivalent 2d image.

By the way, unless you have camera array arranged in a circular fashion, you won't be shooting true, rotatable 3d anyways.
CSharpner
5 / 5 (2) Dec 24, 2010
Evgenijm,
There is a limit to how much memory you need for a given task. Even tasks, such as video recording have an upper limit. For example - it is meaningless to record a movie at resolution which is higher than "resolution" of human eye. Eventually we will have more memory then we need to use for our task and goals. Heck even now I ask myself what to do with 350 gb of my 1 tb hdd.

I could definitely use video that's orders of magnitude higher resolution than the human eye. I could zoom in to get more detail of a scene. zooming in has a whole plethora of uses.

and again, 3D holography has enormous size requirements. we haven't even begun to touch on AI yet, which will easily consume everything we through at it.
Quantum_Conundrum
not rated yet Dec 24, 2010
Now picture if you had an orientable molecule at the nano-scale, such that the body of the molecule is symetrical, except perhaps one atom in an orientable direction. Then you can imagine that it might be possible to manipulate this molecule, making some sort of gear or cog, so that it works just like the hand on a clock. Then you can store "base n" on the molecule based on what fraction of a turn it has experienced: 12 oclock, 1 oclock, 2 oclock....11 oclock, etc.

In theory, it should be possible to store at least one number in base 12 on a single molecule through a mechanical means, by turning a nano-motor or manipulating a magnet or laser.

Then assuming the nano-motors or laser, etc, isn't larger than it's worth, you should end up with greater data density than anything I've seen so far. However, it will certainly be much slower than much of anything too, because it would require a device to read it and convert it to binary and vice versa.
Quantum_Conundrum
not rated yet Dec 24, 2010
By the way, unless you have camera array arranged in a circular fashion, you won't be shooting true, rotatable 3d anyways.


I agree. But hollywood and video game developers have access to the funds to do those types of projects using (presumably soon,) combination phased array optical cameras and lidar, you could probably do a "true" hologram with around 5 to 10 camera stations in the following locations with respect to subject/origin:

+/- x
+/- y
+ z
4 diagonals

So front/back, left/right, overhead, and 4 diagonal overheads using phased optical array detectors should eventually allow accurate, "better than human eye" quality 3-d mapping of just about anything.
CSharpner
5 / 5 (1) Dec 24, 2010
QC,
So a hologram has thousands of times more pixels, and each pixel requires at least 2 more bytes of storage space just to record it's coordinate

The premise of your response it true; we do need more storage for 3D, but your understanding of how we store pixel data is incorrect. in the largest image files(uncompressed), we never store pixel location... only color data about the pixel. its coordinates are infered by its location in the file. the size of the image (width, high, color depth) are usually stored near the beginning of the file, then the rest is just a 2D array of colors. for 3D, we'd also include depth size in the header and our pixel data would be a 3D array.

but your overall point is true. 3D takes more space.
Quantum_Conundrum
not rated yet Dec 24, 2010
Alternatively, you might store higher bases in a single molecule via varying the temperature of the molecule.

Have it set up where distinct values are seperated sufficiently on the temerpature scale to be well outside the margin of error.

Then let's say you can safely vary the temperature of the molecule by 30 degrees, and the margin of error in your measurement is is like 0.001 degrees. Then you should be able to easily store data by incrementing the temperature of the molecule by some easily maintainable metric, such as 0.1 degrees. This would means you could store numbers 0 to 299, more than a byte, in a single molecule if you can vary the temperature across 30 degrees.

Ok, so this is probably an exaggeration, but it is another example of trying to think outside the box of "bits and "bytes," electrons and magnetism.
Quantum_Conundrum
not rated yet Dec 24, 2010
Ok, the temperature thing wouldn't work at all for long term storage, but my be useful in some other way like RAM. It's just an example of a theoretical way to use something other than electricity or magnetism to do math or in this case temporary storage.

Like instead of temperature, spin the molecule and count how fast it's spinning. each rpm could equal 1 increment in a non-binary numbernig system....
CSharpner
5 / 5 (1) Dec 24, 2010
Dtxx,
Plus in this case the shape is easily represented as a vector.

While true, we generally don't vectorize scanned images. Especially not photos. We use different algorithms that approximate the original pixels. This is not nearly as space efficient as "knowing" the true bounds of an object, but is significantly more efficient than no compression. 3D, however, will improve upon this because of the nature of 3D, we'll have better object boundary information available to make /some/ border identification, but lots of it will still have to be stored in a 3D pixel approximation. That's for STORAGE, but for display, 3D array of pixels, similarly to how your monitor and tv have a full 2D array of uncompressed pixels, even when displaying a compressed photo.
Quantum_Conundrum
not rated yet Dec 24, 2010
The problem with most alternatives other than the "nano-clockwork" idea is that they suck for long term and they suck for mobile devices.

Although I thought of one other theoretical possibility, which is of course to literally count atoms like counting marbles or jelly beans in a jar.
Quantum_Conundrum
not rated yet Dec 24, 2010
Ok, here's a theoretical "non-binary" way to store data in extremely dense manner. Again, we assume an intermediary "conversion" step for a computer to use the data.

Consider we have a "stick" made of a single Carbon Nanotube, then we can store an half-Byte of data in that nano-tube physically by moving the tube left to right by a distance equal to the width of one carbon atom, and clamping it in place. The length on the one side corresponds linearly to the value we are storing, 0 to 15.

A nano-tube is about 12nm diameter, and you'd only need length equal to about 16 atom's width, plus however much you need for a clamp of the mechanism to move it. This would theoretically allow 9 half-bytes stored in an area about 39nm by 39nm, and about 20nm tall. Compared to the best cache in a 32nm process chip, which is currently just one BIT in a 32nm by 32nm transistor, not even counting wires...

So this could theoretically store 4.5 bytes mechanically in the same space as a modern bit.
Quantum_Conundrum
not rated yet Dec 24, 2010
Additionally, you might even be able to use a "thinner" molecule, since carbon nanotube is a hollow tube. So maybe a strand of a polymer instead, and measure how much you've pushed/pulled past a reference point.

Even the nano-tube approach could theoretically obtain a much higher data density than DNA, by the time you count DNA bases are doubled, and the deoxyribose or ribose part is about the same as the machinery to move the sticks, and the read head might be a laser or a force microscope analogous to a ribosome. The data density is similar to DNA, if not greater.

If you used the polymer as in paragraph 1, you might be able to obtain data densities as much as 49 nibbles (half bytes) in an area of 36nm by 36nm by ~20nm height.

Even DNA can't be any denser, because polymers are actually narrower than DNA bases, plus again, the DNA has the deoxyribose and the pairing to waste space. If you interpret DNA in codons of 2 = base 16, and not as dense as just using polymer "sticks".
fmfbrestel
not rated yet Dec 24, 2010
To anyone who wants to record their life, home, car and pets from 16 angles at once with 3d holographic cameras: You better start making one damned fine visual search algorithm, and start making it now. Or hire an army of security forces to monitor all of your video feeds.

Technology is limited by our capacity to gain useful benefit from deploying it.
fmfbrestel
not rated yet Dec 24, 2010
For example: Google wave is a highly sophisticated technical innovation, but no one uses it because it doesn't really help them.

However regardless of consumer uptake, if you offered Google Petabyte memory drives, they would probably buy out your entire production capacity.
fmfbrestel
not rated yet Dec 24, 2010
could you imagine Street View in 3D holographic accessed from you phone?

Quantum_Conundrum
not rated yet Dec 24, 2010
To anyone who wants to record their life, home, car and pets from 16 angles at once with 3d holographic cameras: You better start making one damned fine visual search algorithm, and start making it now.


Actually, 3d will make the visual search algorithm easier, because these devices will work just like Radar and Sonar, except with visible and infrared light. This will allow the computer to find "boundaries" much easier than just using two dimensional photography. The software will be able to cross reference the infrared, visible, and lidar aspects to spot exactly where boundaries are and exactly what their colors, temperatures, and shapes are in 3d.

Whereas conventional vision finding software works by contrasting pixels and looking for specified two-dimensional shapes, and they aren't very good at dealing with it when the shapes aren't oriented precisely. With the tools I'm talking about, this won't be a problem.
Quantum_Conundrum
not rated yet Dec 24, 2010
I figure a camera marketable for personal use will eventually be able to map your face from arm's length (self portrait,) accurate to within about 1 micron in each dimension.

A hollywood studio or game developer using a system of cameras to see all surfaces, and correct for one another's margins of error, should be able to eventually map the 3-d surface of an actor's body accurate to within the wavelength of visible light.
fmfbrestel
not rated yet Dec 24, 2010
OK, yeah, that sounds like a great way to do it, but it also sounds like about 2 million lines of code. Now, if that code becomes public standard you could easily still make it happen.

I worry about an FBI surveillance program with Street View like database with constant updates. I am very libertarian when it comes to police actions, it might be about time to update our privacy laws, IMHO.
Quantum_Conundrum
not rated yet Dec 24, 2010
See, what they can do is for the lidar aspect, send out two pulses, one in infrared, and one in violet light. The infrared is invisible to the end viewer anyway. The violet can be pallete-swapped at editing time or "run time" on a video game. this is how player color is done in video games anyway.

These pulses are used the same way radar and sonar works, to measure the surface of the object/person, giving an accurate 3-d return signal, who's resolution is limited only by the wavelength of light and the scale of the detectors, which would be a phased array of nano-dipoles (probably carbon nano tubes or something similar.)

So for science and forensic cameras, they can have resolution approaching nearly 300,000 times what most human eyes can discern at arm's length...in each dimension.
fmfbrestel
not rated yet Dec 24, 2010
The Singularity is Near.
fmfbrestel
not rated yet Dec 24, 2010
Already we are developing technologies that we cannot even fully utilize yet, and the pace of advancement does not seem to be slowing. Because those most able to take full advantage of these advances are the people making the advances.
Quantum_Conundrum
not rated yet Dec 24, 2010
Already we are developing technologies that we cannot even fully utilize yet, and the pace of advancement does not seem to be slowing. Because those most able to take full advantage of these advances are the people making the advances.


yeah, well, I find people tend not to realize what new technologies are capable of. We have these smart phones and stuff, but we aren't realizing how world changing this technology can be.

We have the computer technology to do self-driving cars now, and yet only a handful of groups in the world are even working on the actual application of these technologies. All the individual component technologies exist.

Then there's nanotech. The applications in medicine and manufacturing are almost fantastic, and yet the government has spent very little for R&D. We should be spending as much or more in nanotech than anything else.

When you try to explain what this stuff definitely can do, and what it probably can do, people just laugh or shrug..
fmfbrestel
not rated yet Dec 24, 2010
A distracted content citizen is much easier ruled then a focused discontented citizen. Just ask Iraq.
CHollman82
not rated yet Dec 28, 2010
Well, first off, there wont be petabyte drives in PC's because there is no need for that much space.


Are you serious?

I have 4 terabytes filled in my home PC... I could easily fill much much more also. I am considering buying 2 more 2tb drives instead of deleting some of my bluray videos and 1080p television shows...

As soon as 3D takes off, which is just around the corner (next few years at least), you will see 100gb+ 3D movies. 10 movies per terabyte... I have over 100 movies on my hard drives and god knows how many hundreds of hours of television shows right now.

I don't know what you use your home PC for, but don't dictate what I can or should be able to use mine for, and don't make stupid statements like "There is no need for X amount of memory/hard drive space"... because you will proven wrong every single time.

(and that's just home PC's, forget about enterprise systems that ALREADY house petabyte databases...)
desotojohn
not rated yet Dec 30, 2010
Sounds like Magnetic Bubble Memory on steroids ...

More news stories

Atom probe assisted dating of oldest piece of earth

(Phys.org) —It's a scientific axiom: big claims require extra-solid evidence. So there were skeptics in 2001 when University of Wisconsin-Madison geoscience professor John Valley dated an ancient crystal ...