backtop

Print E-mail del.icio.us 34 comment(s) - last by zephyrprime.. on Jul 3 at 11:28 AM
Lab experiments show laser hard drives 100x faster than conventional technology

When it comes to computer technology, hard drives have advanced the slowest over the years are far as speed is concerned. We've seen dramatic increases in processing speed and graphics power, but HDDs have long been the bottleneck in modern systems.

Many are looking to solid state disks (SSDs) to give a boost in speed. In addition, SSDs have the advantage of lighter weight, more efficient packaging, silent operation, durability and power efficiency.

The major drawback, however, is pricing. Current 1.8" SSDs are roughly five times as expensive as their 1.8" HDD counterparts. Analysts expect the difference to only drop to three times as expensive by 2010.

With this in mind, researchers are looking to other alternatives to traditional HDD technology. Researchers at Radboud University Nijmegen are using lasers to write data to HDDs. Each laser pulse is capable of heating up a portion of the magnetic disk (made up of cobalt, gadolinium and iron) to change its polarity. By doing this, 1's and 0's can be recorded at rapid speeds.

Researchers claim that the use of lasers makes the drive 100 times faster than conventional hard drives.

"This is one of the most exciting stories in magnetics," remarked Julius Hohlfeld of Seagate Research. While current laboratory experiments have been successful, a working prototype isn't expected for another 5+ years.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Hmmm...
By DeepBlue1975 on 7/2/2007 10:01:38 AM , Rating: 2
This talks about writing. What about the reading part?
In which way are they going to be 100 times faster, in raw sequential speed, or in access times, or both?

I guess Patentman could give us some insight here...
If this really can make a performance difference, maybe it could fill a transitional gap between a conventional HDD and an SSD, until the latter can start having a good price / storage ratio which can bring it to mass adoption.




RE: Hmmm...
By noxipoo on 7/2/2007 10:11:51 AM , Rating: 2
forget SSDs if this is 100x faster overall, or even 50x faster overall.


RE: Hmmm...
By peldor on 7/2/2007 10:18:04 AM , Rating: 4
100x faster sounds great, but the access time of 5 years is a bit high. You'll probably have the SSD first.


RE: Hmmm...
By TimberJon on 7/2/2007 11:05:55 AM , Rating: 2
DeepBlue1975's comment and question is valid.

Whoohoo if you can write 100 times faster, but if you can only read at 10 times faster, while that is still an improvement, it will be looked down upon. OR they just won't release such a product until they can get write and read times closer together.

I'd like a drive that could write even just 20x faster. Better for backup purposes. And the way big companies are storing data these days, (I.E., Yahoo and unlimited storage) we need faster backup solutions yesterday. Yea they have redundancy, but you can never have enough backup.

Ask a Marine.


RE: Hmmm...
By theapparition on 7/2/2007 12:15:44 PM , Rating: 3
I've never seen any medium that could write faster than read. Someone correct me if I'm wrong.


RE: Hmmm...
By AsicsNow on 7/2/2007 3:09:04 PM , Rating: 2
Who gives a damn because its still not gonna fill in the same market niche that SSDs are gonna occupy. It will still need a motor, AND a laser, thus one of the biggest selling points for SSDs will remain valid: battery life .

Could be pretty damn nice for desktops and well laptops that don't care as much about gaining the energy efficiency that SSDs offer.

However, the laptop/mobile market is going to become even more dominant before this technology is even close to release, and unless some miracle chemist cracks the secret to creating dilithium, there probably aren't going to be too many big jumps in the technology behind batterys.


RE: Hmmm...
By thatguy39 on 7/3/2007 12:26:44 AM , Rating: 2
quote:
Whoohoo if you can write 100 times faster, but if you can only read at 10 times faster, while that is still an improvement, it will be looked down upon.


uhhhh... WHAT?

you mean to tell me a HD that can read 100x faster and write 20x faster is gonna be looked down upon by you?

Please explain to everyone what it is your doing that requires write speeds to be 20x what they are today... please, tell us.

You make no sense, Burners came out at 1x... 52x read speed, 1x write.

Come on buddy...


RE: Hmmm...
By TomZ on 7/2/2007 11:14:04 AM , Rating: 1
From what I can understand of this research, I would think that access time is still going to be a limiting factor. I would guess the 10-100x being discussed applies to read/write transfer rates, but seek time is still going to be a function of the rotational latency and therefore rotational speed.


RE: Hmmm...
By masher2 (blog) on 7/2/2007 11:15:32 AM , Rating: 5
> "In which way are they going to be 100 times faster, in raw sequential speed, or in access times, or both?"

In writing times. Access times are governed primarily by platter spin rate and head positioning delay, neither of which would likely be affected by changing the write head. Of course, one could always postulate a different means for positioning the laser...something that isn't really practical for a magnetic head.

Also, I don't see how this would be applicable to replacing the read head, so don't look for any dramatic advances there.


RE: Hmmm...
By drank12quartsstrohsbeer on 7/2/2007 4:50:19 PM , Rating: 2
I have often wondered why drive manufacturers never made drives with more than one actuator arm (or whatever its called).

Sure it would take more processing power and energy to control and reconstruct the data. But I am sure there are enough people willing to pay the premium for it. The WD raptor is a perfect example.

And what ever happened to those TrueX CD roms, that had 7 lasers instead of one?


RE: Hmmm...
By plewis00 on 7/2/2007 8:15:20 PM , Rating: 2
You're not the only one - I thought two arms in one drive would potentially be a lot faster. Like you say there's more data processing involved but still... It could even have 2 data connectors, so the machine treats and sees it like a RAID0 device, it could be totally seamless to the computer itself - though physically it is one set of platters, and two heads.

It's like processors moving to dual and quad-cores - everything's moving towards parallel-processing (apart from serial data transfer). It would appear to make sense. I guess there is a reason for this - more parts to fail maybe? I would've thought cost would be less of a consideration as like you say Raptors have been snapped up.

The TrueX CD drive was conceived and built by Kenwood but it failed due to complexity and cost. It wasn't 7 discrete lasers it was one laser with lenses to diffract the light, very complex and from the reviews I read, on the right media it was fast but on lower-grade media (read cheap CD-Rs) it was terrible, I suspect because of alignment reasons - standards drives are infinitely more tolerant.


MO?
By Slaimus on 7/2/2007 11:03:41 AM , Rating: 3
How is this different than the MO drives that have been out for a while? Even MiniDiscs are just small MO drives, which are magnetic but written with lasers.




RE: MO?
By masher2 (blog) on 7/2/2007 11:29:04 AM , Rating: 2
A MO drive still uses a magnet for writing the data. The laser is used for reading, and to heat domains on the platter during writing.


RE: MO?
By highlandsun on 7/2/2007 4:07:52 PM , Rating: 2
Right, minidiscs use a laser to heat a spot on the disc to its Curie point. A magnet is then used to write the 1 or 0. When the laser/heat is removed the spot cools down and the bit is essentially frozen into the disc surface. MD data is extremely persistent; without the intense heat it takes a fairly powerful magnet to wipe out a signal recorded this way. Of course MDs are pretty slow...

The other points about seek times and rotational latency are still significant. But it's possible that a micromirror array (like TI DMD) could be used to eliminate those delays. In that case you no longer need a rotating platter, just a square of material will do.


100x faster bit writes != 100x faster drives
By jrk on 7/2/2007 11:59:01 AM , Rating: 5
This research is interesting and important, in the long run, but just to be clear: all they are doing is decreasing the time it takes to force a state transition in one of the magnetic bits on the platter by 100x. This just means the bit need only be exposed to the head for 1/100th the time to produce an effective write. For this technology to bring about 100x faster sequential writes would require spinning the platter 100x faster -- that is the main thing this would allow you to do in current technology. Going forward, as magnetic bits shrink, they become easier to flip, though, so even this advantage will shrink dramatically in the "5 years" it will take to commercialize this technology.

Where this is probably interesting, though, is in enabling faster writes to more state-change-resistive materials. As bits get smaller (and proportionally easier to flip, in a given design) it will be necessary to use materials which are more magnetically stable. The problem this creates is that they will not only be harder to spontaneously flip (creating an error), they will also be harder to write. Heat-assisted writes of this sort could enable fast writing to very small bits in materials which are otherwise magnetically stable. That is why this technology is interesting: because it is one piece of the broader effort to create ever-denser magnetic media.

And all of this, of course, says absolutely nothing about latency/access time or even sequential read bandwidth. This does not spell the end of SSDs. Indeed, it is largely orthogonal.

Cf. http://storagemojo.com/?p=487 for more.




By TheRodent on 7/2/2007 12:32:12 PM , Rating: 2
I think in terms of potential SSDs spell the end to all moving hard drive storage, laser writes or otherwise. This advance is too little too late. The limitations and costs of a physical system to place the read/ write head will eventually make moving hard drive advances impractical. Until this point the ability to move a single head in two dimenions has allowed cheaper storage than available with semiconductor technologies. This will inevitably change with the falling cost of Flash devices. Let alone advances such as phase change memories.


It is a fair day, today.
By NuclearDelta on 7/2/2007 1:06:36 PM , Rating: 2
I see about writing to the platers in the article, but I have yet to see the words "Faraday rotation", which I assumed they would use to read the data back.




RE: It is a fair day, today.
By masher2 (blog) on 7/2/2007 1:44:53 PM , Rating: 2
Most likely it would be the Kerr effect, actually, which relates to light reflecting off a magnetized material. The (very similar) Faraday effect relates to light passing through a magnetized medium.


RE: It is a fair day, today.
By NuclearDelta on 7/2/2007 1:57:03 PM , Rating: 2
Aye. I appears I was thinking one and named another. Thanks for the correction, though.

I wonder if there is a cycling limit to the platter material, and what the temperature threshold of the "flip" is. I am also curious to what degree(to extent and temp, ba-zing) the localized heating, for long write periods, will increase drive temp compared to traditional drives.


It's obvious.....
By marvdmartian on 7/2/2007 10:10:23 AM , Rating: 5
.....that I've been hanging around here too long! When I read the link description about laser hard drives, the first thing that came to mind was,
quote:
"All I want is some frikkin' hard drives with laser beams!!"


then I open the page, and the first thing I see is Dr Evil!! LOL




Reminds me of Zip Disks
By mmcdonalataocdotgov on 7/2/2007 10:28:57 AM , Rating: 2
Zip disks filled in between floppies and CDRW, and where are they now? Like the Bob interface between Windows 3 and Windows 95. Where are they now?




RE: Reminds me of Zip Disks
By arazok on 7/2/2007 10:14:34 PM , Rating: 2
Zip drives failed (IMO) because Iomega kept them proprietary, and expensive. If Iomega wasn't so greedy with that technology, I think it could have been a bigger format then it was.


Dr. Evil
By corduroygt on 7/2/2007 2:13:18 PM , Rating: 2
I knew before clicking on the link that there would be a Dr. Evil picture, their sense of humor is one of the reasons I love DT.
Now when can we get some hard drives with fricking laser beams eh?




RE: Dr. Evil
By Bladen on 7/3/2007 4:51:18 AM , Rating: 2
Let's just hope that these hard drives aren't mutated or ill-tempered.


5 years
By knowom on 7/2/2007 4:17:33 PM , Rating: 2
In the five years it takes for them to have a working prototype let alone anything to sell to the market SSD's will be cheaper, faster, more durable, and likely better on battery life it's got better access times and is silent. SSD is the future at this point I mean in 3 years it'll be only 3 times the costs of traditional HD's so withen 5 it should be somewhere within 1.5 to 2 times the cost while being significantly better overall. What you pay more will be partially offset by what your paying extra in power consumption plus they run much much cooler.




RE: 5 years
By penter on 7/3/2007 7:05:55 AM , Rating: 2
Yeah, and in the meantime, harddrives will just sit still!


By lynxss on 7/2/2007 2:28:20 PM , Rating: 3
Does anyone remember the optical based drives from the late 90s that operated off of a pair of lasers triangulating in a 3 dementional space inside a optically sensitive crystal type material to set and read 1/0s.

In the prototype they were able to store something like 250 full length moves several terabytes worth of data in a 1inch X 1inch X 1inch cube of the stuff. This was close to a decade ago and we still have nothing that even comes close to that type of storage density.

What happened? I remembered it being fast enough to play multiple streams of video simultaniously but they were still working on some bugs, then nothing else on it for many years.




About speeds
By DeepBlue1975 on 7/2/2007 3:55:11 PM , Rating: 2
Masher2:

That's what I suspected. This only talks about STRs. Even though I've read the linked article, it says nothing about the way this is going to be implemented.
When I read "laser" I thought that, just maybe, the heads could be than in a way comparable to the one used on those old kenwood true X optical drives, which could potentially lead to better access times.
If we're just talking about improving STRs and just replacing the magnetoresistive heads for laser ones while keeping the rest of the structure (that is: a single motorized arm which travels across the disk's radius), I wouldn't be much excited about this.
100 hundred times de STR with the same access times, how much of a "real world" performance improvement could give us, desktop users?

No doubt this could be great for streaming applications, but as for normal desktop access patterns, I don't know if it would be so much of an improvement.
That and the fact that having a working prototype "within a decade" seems to be to big of a time window, maybe big enough to allow SSDs to get capacities and prices much closer to HDDs.




By Reegor on 7/2/2007 5:00:03 PM , Rating: 2
Thermally assisted magnetic drives have been under research for years. (e.g. http://news.com.com/A+divide+over+the+future+of+ha...
This article claims a whole different mechanism, but remember that the size of the smallest laser spot is considerably larger than current magnetic domains. (Even blue laser has wavelength of .5 micron or 500 nm) So density of storage, which is the main constraint on reading and writing speeds, would be lower. The article claims they will eventually reach 10nm , but there is no mass-produced laser even remotely capable of such a wavelength.
quote:
One drawback is that the footprint of the laser pulse on the disk is about 5 microns wide, which is considerably larger than the footprint produced by existing data-transfer systems. But physics doctoral candidate and co-author Daniel Stanciu says the team is working on improvements in the technology that should reduce the footprint's size to about 10 nanometers, and he expects to see a working prototype within a decade.

So, this is (yet another) instance of a mechanism that theoretically, in some alternate universe, could beat standard magnetic recording, but in practice won't come close to keeping up with the ~50% per year improvements in areal density.

As for solid state disks, their ultimate triumph has been predicted at least since the 1970s. For all of that time, though, magnetics have gotten cheaper faster than solid state. However, SSDs have a great future for small devices. The crossover size (at which advantages of SSDs outweigh their higher cost) is currently 10 to 20GB, and will continue to rise.




By zephyrprime on 7/3/2007 11:28:21 AM , Rating: 2
The problem is, photons are too big. The data density of such a device would be inadequate.




Another punch for my nitpick card
By peldor on 7/2/07, Rating: -1
RE: Another punch for my nitpick card
By masher2 (blog) on 7/2/2007 11:20:56 AM , Rating: 5
> "They are Solid State Drives ...

A word means what we intend when we use it. Few words are truly logical if one examines their etymology. Disc drives were named because a "disc" was being rotationally "driven". While a SSD has no disc in it, its also technically not a "drive" either.

Luckily, we need not be so pedantic. Call it a disc, a drive, or a solid state storage device...we'll still know what you mean.


RE: Another punch for my nitpick card
By DigitalFreak on 7/2/2007 11:31:01 AM , Rating: 5
He was just being a dic.


By natebsi on 7/2/2007 12:55:01 PM , Rating: 1
Don't get clever on us!


"So if you want to save the planet, feel free to drive your Hummer. Just avoid the drive thru line at McDonalds." -- Michael Asher

DailyTech Poll
Which web browser do you use on your primary personal machine? 






44 Comments









botimage
Copyright 2009 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki