Friday, November 26th 2010

Hitachi and Partners Develop New HDD Technology Providing 8-Fold Density Increase

A consortium led by Hitachi, including Japan's New Energy and Industrial Technology Development Organization (NEDO), a public-sector body that promotes research of energy-efficient technologies, National University Corporation Tokyo Institute of Technology, and Kyoto University, have developed a new hard drive data recording technology that promises to increase data density 8-fold.

The researchers have developed a new patterning technology for the magnetized bits that are laid on the platters, which are just 10 nm in size. This is made possible by using new materials, and making use of a self-arranged phenomenon of polymer materials. Its practical applications increases areal-density of disk platters to 3.9 Terabits per square inch, an 8-fold increase compared to the 500 Gigabits per square inch that's currently standard. This paves the way for 24 Terabyte (TB) hard drives in the very near future.
Source: CDRInfo
Add your own comment

51 Comments on Hitachi and Partners Develop New HDD Technology Providing 8-Fold Density Increase

#27
Completely Bonkers
I am really looking forward to these. I have 1UE rack's running 2.5" drives. Nice, small, cool, quiet. But currently with just two drive bays I'm limited to 2x640GB, and if I RAID 1 then just 640GB. While I am OK with that storage space today (we only use 30% of it on documents, PDFs, etc) it will *not* be enough next year when we intend to store video material too.

Even if we don't go video, then having a SINGLE HDD backup server that can capture everything in the office would be marvelous! There is nothing more time-consuming that managing multiple drives with different backup hypothecations. One drive, snap the office, job done :)

So long as data integrity is maintained with these new drives, GREAT!
theJesusDiamonds are the hardest metal known to man.


Diamonds are not metals.
Posted on Reply
#28
theJesus
Completely Bonkers

Diamonds are not metals.
Know your meme :slap:

Anyway, I will definitely be getting 24TB drives when they're available :rockout:
Posted on Reply
#29
Swamp Monster
theJesusKnow your meme
FYI In Your Link:
Because diamond is an insulator rather than a conductor of electricity, it is not classified as a metal.
Posted on Reply
#30
theJesus
Swamp MonsterFYI In Your Link:
Because diamond is an insulator rather than a conductor of electricity, it is not classified as a metal.
I think you're missing the point :rolleyes:

Back to the HDDs please?
Posted on Reply
#31
Swamp Monster
Of course, I hope these new HDD's will be avaliable in very near Future, and prices will not make my wallet empty.
Posted on Reply
#32
theJesus
I anticipate that I'll be done college by the time the 24TB drives are out, so I'm definitely getting 6 :D
Posted on Reply
#33
Completely Bonkers
theJesusKnow your meme
Crickey! Proof that there is a reason we should censor the web! We should issue "licenses" to allow some people to WRITE and POST material! ;) Amazing how we let these things develop. Someone says something stupid, and then it becomes cool to quote this disinformation, and some poor sucker who is hoping to learn, picks up the proverbial wrong end of the stick.

Personally, I disagree with lowering educational and scientific standards to the lowest common denominator, and allowing illeducated memes dominate, confuse, or muddy our thirst for knowledge. (when they are wrong). There are plenty of chat/nonsense sites full of lax editorial and rubbish. Let's stay away from them, and not let them near our beloved TPU. :)
Posted on Reply
#35
AsRock
TPU addict
Very nice and i am sure price be high for a while until others start doing ones with such high density.
mechtechHmmm 8 fold increase in density!! I wonder what kind of real world increase in speed it would gain? 3 fold maybe?

If so they could read/write around 300MB/s which would make ssd almost obsolete if not for the quick access/load times.
It's the sata port whats slowing SSD's down if there was higher speed SATA there would be higher speed SSD's.

And they could do faster SATA ports and controllers but it's all about making money like hell SSD's dont get hot yet :P.. Worry about SSD's when they have massive heatsinks on them ha.
Posted on Reply
#36
LAN_deRf_HA
Here's an idea. I don't need more than 2 tbs. I get a 12-24 tb drive. If I short stroke that down to 2 tbs I should get a significant boost in access time, and better average transfer rates. Still no ssd, but maybe raptor level or lower at only 7200 rpm. An actual raptor using this and short stroked might get down to 4 ms. Right now it's only a neat trick, but if we hit a truly absurd over abundance of storage space I could even imagine a line of performance drives shipping short stroked from the factory.
Posted on Reply
#37
TheLaughingMan
LAN_deRf_HAHere's an idea. I don't need more than 2 tbs. I get a 12-24 tb drive. If I short stroke that down to 2 tbs I should get a significant boost in access time, and better average transfer rates. Still no ssd, but maybe raptor level or lower at only 7200 rpm. An actual raptor using this and short stroked might get down to 4 ms. Right now it's only a neat trick, but if we hit a truly absurd over abundance of storage space I could even imagine a line of performance drives shipping short stroked from the factory.
You can make them small so you can run them at high speed. With this density we can get 1 TB 15000 RPM Raptor drives. Or 2 TB drives in Laptops, super small externals (like 1 and 2 TB drives that are only slightly larger than thumb drives).

This could also help make the Retangular Hard Disk drives more viable since we can keep the current drive sizes, but make the RHD's the size of SSD's.
Posted on Reply
#38
Kantastic
TheLaughingManThis could also help make the Retangular Hard Disk drives more viable since we can keep the current drive sizes, but make the RHD's the size of SSD's.
Wonder what ever happened to the develop of RHD's.
Posted on Reply
#39
TheLaughingMan
KantasticWonder what ever happened to the develop of RHD's.
The tech is still be worked out. Their are heat issues and ware issues related to it vibrating in stead of spinning. The last I checked on it, one of the devs. wanted to suspend the internal parts in a liquid. Then everyone got freaked out about it.
Posted on Reply
#40
Steevo
I need more room, 3TB with only a few hundred gig left, and I haven't backed up my movies, video, or installed half my games.


I foresee the need for a 10TB server in my immediate future. On Gig network of course.
Posted on Reply
#41
dr emulator (madmax)
Musselshell no, i've got about 15TB of data and NEED MOAR.
:laugh: i've got that many drives i've actually know idea how many drives i have got :twitch::wtf::D

actually just think of all the hot gals you could store, mind blowing:twitch::pimp:
Posted on Reply
#42
WarEagleAU
Bird of Prey
Wow, very awesome. Now if they can get them closer to SSD speeds....lol.
Posted on Reply
#43
LAN_deRf_HA
For the average person 10% of a 24 tb drive would be enough. Look at the access time on a 10% shortstroked 2tb. Raptor territory.

Posted on Reply
#44
Wile E
Power User
xairaanybody else think we should slow down, i mean come on, wasnt it just the other day we were astounded by the intro of 1tb, now 3tb, very near to 10tb yeh, but24 :O_o:
No. Not ever. There is never enough storage or speed, EVER.

I want at least 6 of these 24TB drives to put in my server in a RAID 5. That should last me at least until the 100TB or possibly even the 1PB drives release.
Posted on Reply
#45
TheLaughingMan
LAN_deRf_HAFor the average person 10% of a 24 tb drive would be enough. Look at the access time on a 10% shortstroked 2tb. Raptor territory.

i200.photobucket.com/albums/aa114/xtremetiramisu/Overclocking/black2tb-b.png
It doesn't quite work that way. You are not accounting for the fact that a single accessed sector would contain 8 times the information. In other words, Arel density directly affects throughput as well and storage capacity. Assuming nothing else changes about the drives, seek time would improve some, but once it found the data we are talking about a difference in the amount of available data. A single revolution of a Raptor drive, short stroked or not, will contain X amount of data. The same single revolution of the new density system will contain 8X data. This drastically alters the MB/s.

You can also look at it in this way. Lets say we have a Blu-ray disc (50 GBs) and a DVD (4.7 GB) in identical drive bays, but running at 8x read speed. When then program the system to read 1 GB's of data off the disc and into the system. Assuming both system are exactly the same, the DVD will on have to spin completely around to almost 20% of the disc's surface; however, the Blu-ray will only have to spin around to about 2%. Since they are spinning at the same speed, the Blu-ray will have completed its read long before the DVD has.

This improvement is almost a perfect parallel to Blu-ray except with HDD the systems connected to it have already been proven to be able to handle the increase in data throughput. SATA 6Gb/s is ready and waiting, and Burst Read/Write speeds show HDD's other components to be able to handle more data. While it will not yield a perfect 8 fold improvement in throughput, I personal don't think a 4 or 5 fold throughput improvement is out of the question.

In the end, the only concerns from our stand point is A. the cost of this new polymer and B. time needed to go from where they are now to mass production in place.
Posted on Reply
#46
zAAm
TheLaughingManIt doesn't quite work that way. You are not accounting for the fact that a single accessed sector would contain 8 times the information. In other words, Arel density directly affects throughput as well and storage capacity. Assuming nothing else changes about the drives, seek time would improve some, but once it found the data we are talking about a difference in the amount of available data. A single revolution of a Raptor drive, short stroked or not, will contain X amount of data. The same single revolution of the new density system will contain 8X data. This drastically alters the MB/s.

You can also look at it in this way. Lets say we have a Blu-ray disc (50 GBs) and a DVD (4.7 GB) in identical drive bays, but running at 8x read speed. When then program the system to read 1 GB's of data off the disc and into the system. Assuming both system are exactly the same, the DVD will on have to spin completely around to almost 20% of the disc's surface; however, the Blu-ray will only have to spin around to about 2%. Since they are spinning at the same speed, the Blu-ray will have completed its read long before the DVD has.

This improvement is almost a perfect parallel to Blu-ray except with HDD the systems connected to it have already been proven to be able to handle the increase in data throughput. SATA 6Gb/s is ready and waiting, and Burst Read/Write speeds show HDD's other components to be able to handle more data. While it will not yield a perfect 8 fold improvement in throughput, I personal don't think a 4 or 5 fold throughput improvement is out of the question.

In the end, the only concerns from our stand point is A. the cost of this new polymer and B. time needed to go from where they are now to mass production in place.
He might be implying that the latency on short stroked drives are pretty low as it is (as in his example). Then when you short stroke a 24TB at 10%, you'd get 2.4TB of space with comparable latencies, although the speeds will be significantly higher of course due to the massive increase in areal data density.

So then you'd have a high speed 2.4TB drive with decent latencies (almost comparable to a Raptor probably, given a constant spindle speed again). I'm guessing he meant that such a drive would be plenty for most users. ;)

Of course, short stroking a 24TB to 200GB would maybe give a slight advantage in latencies due to the head only having to move slightly side-to-side to access any given sector, but the limiting factor would always be the spindle speed as you explained in excruciating detail. :laugh:
Posted on Reply
#47
TheLaughingMan
zAAmHe might be implying that the latency on short stroked drives are pretty low as it is (as in his example). Then when you short stroke a 24TB at 10%, you'd get 2.4TB of space with comparable latencies, although the speeds will be significantly higher of course due to the massive increase in areal data density.

So then you'd have a high speed 2.4TB drive with decent latencies (almost comparable to a Raptor probably, given a constant spindle speed again). I'm guessing he meant that such a drive would be plenty for most users. ;)

Of course, short stroking a 24TB to 200GB would maybe give a slight advantage in latencies due to the head only having to move slightly side-to-side to access any given sector, but the limiting factor would always be the spindle speed as you explained in excruciating detail. :laugh:
Thanks, I just realized that he was talking about access time and not read/write speeds. I have to start doing all this at reasonable times.
Posted on Reply
#48
BorgOvermind
Don't worry ... we won't see this tech on the market any time soon. It takes years to develop it up to mass-production status.
Posted on Reply
#49
MikeMurphy
They need to develop something better than a read / write heads that moves around. Maybe a head strip across the platter.
Posted on Reply
#50
Mussels
Freshwater Moderator
MikeMurphyThey need to develop something better than a read / write heads that moves around. Maybe a head strip across the platter.
look into HRD's, we've already been discussing them in this thread.
Posted on Reply
Add your own comment
Dec 27th, 2024 13:45 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts