Heat Key To Faster HDDs...Hundreds of Times Faster

[btarunr]

Physicists have discovered a new method of recording data on hard drives that could potentially make mechanical hard drives hundreds of times faster (in terms of performance). Heat holds the key. A hard disk drive is a magnetic storage device, which, unlike magnetic tapes, allows random access. Its recording surfaces consist of hundreds of billions of tiny portions that can be magnetized in a particular polar direction to represent 0s or 1s. To create meaningful arrangements of these portions (bytes), the recording heads apply an external magnetic field to flip their polarities.

A team of researchers led by Thomas Ostler at the University of York, UK concluded that the process of flipping the polarities of these tiny portions can be greatly expedited using short bursts of heat. It has been believed that heat could only assist in remagnetization, when used in conjunction with a magnetic field. Research shows that this can also be achieved using a very, very precise amount of heat generated by beaming a fine laser for less than a trillionth of a second, which momentarily raises its temperature by 800 °C. The results of this study was published here.



Researchers say that the new method of magnetic recording can achieve terabytes per second (1000 GB/s) recording/reading speeds. The process also uses less energy than magnetic recording, meaning the new drives would be more energy efficient.

Source: New Scientist, Image Courtesy Gizmodo

[Steevo]

At a few watts for a drive I don't really care about energy use, however I would like them to expound upon the terrabytes per second of reading and writing with their technology, as I am understanding this it uses a laser to write at a much higher speed onto magnetic media, but I'm not sure what they have done differently to make reading any faster?


Grant money hunt anyone?

[m1dg3t]

Interesting, temperature has long been known to affect magnetization but to see it implemented in this manner is a first! Maybe i should pull the fan's off my HDD's

[tigger]

Yaay for British ingenuity.

And very interesting development for hard disc tech.

[WhiteLotus]

Originally Posted by tigger

Yaay for British ingenuity.

And very interesting development for hard disc tech.

Yup just read this, and instantly came over here to see if anyone had picked up on it.

Hurray for the University of York!

[Delta6326]

... How much will cost to add a laser gun to the HD?

Sounds cool(no pun intended) If this get popular than hopefully it will drive prices of more practical stuff down. As it will take a long time to be able transfer stuff at those speeds.

And for the energy use this would be helpful for large servers with TB's of memory. Take a server with over 120TB's say they all are 2TB thats 60HD's that can create a lot of heat and power.

[theonedub]

What interface would support transfer speeds that fast?

[crazyeyesreaper]

PCIe

aka PCIe 3.0 x1 = 2GB/s

[sneekypeet]

Wonder how long it takes someone to buy one of these and strip it down for that 800*C temperature, pew pew laser. Good idea on the tech, but I think letting the average Joe have access to these lasers has to be bad news somehow

[extrasalty]

Can't wait for the sharks with the frickin laser beams.

[m1dg3t]

Pretty sure if you got the cash you can get 1,000mw+ laser's, in a variety of colour's as well! Not sure if they'll hit 800c but they are said to be able to light thing's on fire and cut certain material's

You could have shark's with red/green or blue/purple or any combination lol

[jsfitz54]

My question would be: Will this get defeated by those who want to develope SSDs as the new platform?

Every indication would leave one to believe that mechanical drives are going the way of the dinosaur.

What about Capacity per GB expense?

[Batou1986]

brb benchmarking the raptor in the oven............

[phanbuey]

or... you could buy four modern SSD's now and put them into a 4xRaid 0.

[twicksisted]

Cool idea, wonder how it would be "safely" implemented in home computing.
Temperatures of 800c inside a component as small as an HDD sounds like a fire risk. Also with that heat im guessing there would be a lot of wear & tear.

[a_ump]

Originally Posted by phanbuey

or... you could buy four modern SSD's now and put them into a 4xRaid 0.

sounds like a lot of work compared to if one of these drives were available.

Here's my question, even if the laser only has to beam for a trillionth of a second, and say i'm transfering a massive amount of data(take the server example above) of 120TB, wouldn't the constant beaming cause instability, bc the drive would eventually heat up too much. I suppose that's the only thing holding this back.

[Beertintedgoggles]

Eh, if you look at the numbers it seems all they did was present that the laser only needs less than a trillionth of a second to flip the magnetic polarity so they extrapolated that they could then potentially flip a trillion bits in one second (1 TB, you can see they are using the decimal 1 TB = 1000 GB). While cool that the theory could potentiall lead to faster writes on a mechanical HD, this is still to far off from implementation to get a rise out of me.

Additionally, as was mentioned above, this doesn't appear to help with read times at all.

Edit: Forgot to mention, this isn't going to help seek times either unless they speed up the RPM's.

[Jizzler]

Originally Posted by Beertintedgoggles

Additionally, as was mentioned above, this doesn't appear to help with read times at all.

Edit: Forgot to mention, this isn't going to help seek times either unless they speed up the RPM's.

Combine with: http://www.techpowerup.com/97827/Har...r_to_SSDs.html

Better seek times, more speed

[Sasqui]

Originally Posted by Beertintedgoggles

Eh, if you look at the numbers it seems all they did was present that the laser only needs less than a trillionth of a second to flip the magnetic polarity so they extrapolated that they could then potentially flip a trillion bits in one second (1 TB, you can see they are using the decimal 1 TB = 1000 GB). While cool that the theory could potentiall lead to faster writes on a mechanical HD, this is still to far off from implementation to get a rise out of me.

Additionally, as was mentioned above, this doesn't appear to help with read times at all.

Edit: Forgot to mention, this isn't going to help seek times either unless they speed up the RPM's.

Originally Posted by Jizzler

Combine with: http://www.techpowerup.com/97827/Har...r_to_SSDs.html

Better seek times, more speed

You still have to move either the read/write head, or the media. That'll be the bottleneck (not even counting the current interface technology)

[Kantastic]

This is great and all, but I can't help but feel this is garbage propaganda spewed out by the HDD industry to offset enthusiasm and focus for SSDs. Moving parts will never be as efficient, and when refined, as easily produced as something flash-based.

[white phantom]

Originally Posted by m1dg3t

Interesting, temperature has long been known to affect magnetization but to see it implemented in this manner is a first! Maybe i should pull the fan's off my HDD's

exactly what i was thinking lol all this time a cool drives a happy drive, now they are saying a roasting hot drive is a faster drive

[Steevo]

Cooling your hard drive is a waste anyway, it has been proven that cooler temps on hard drives results in earlier death.

[Shihabyooo]

Well, that would be one OC you won't mind the increased temperature with >_>

[Jizzler]

Originally Posted by Sasqui

You still have to move either the read/write head, or the media. That'll be the bottleneck (not even counting the current interface technology)

Yup, but I'll take it. 64+ heads helps

[Kreij]

Interesting concept but it leaves a lot of questions.
What's the cool down time for the spot the laser hits and cooks to 800°C?
If adjacent heat transfer occurs between laser points, the platter is going to have to be made out of ceramic (think space shuttle tiles) to perform sequential writes.