Storage

Helium-filled drives will lift capacities to new heights

Scott Lowe explores a new frontier in drive capacity -- helium.

Sometimes, physics can be a real pain in the neck. Nowhere is that more true than when it comes to the ever growing demand for newer and larger hard drives. What amazes me is that each time it appears as if the industry has hit a road block when it comes to increasing drive capacity, some really smart people get together and do some amazing things. A few years ago, as storage density began to approach levels that would result in hitting limits as a result of the superparamagnetic effect, the industry just flipped... to perpendicular storage. This staved off the coming magnetic dark ages that would have stopped capacity increases in their tracks.

Today's physics challenge isn't magnetism... its air. Yes, air. If you were to open up a modern hard drive, you would find that it actually has holes in it. As the platters of a hard drive spin, air is also pulled into and pushed out of the drive. The drive's read and write heads ride an extremely thin cushion of air to avoid sliding across and damaging the hard drive platters. Through this tiny air cushion, the drives read and write heads do their duty, reading data from and writing data to the hard drive. However, like a jet engine pulling air into and blowing it out of the engines, this constant flow of air creates turbulence that can push the read and writes heads off their intended track. This results in inefficiencies and makes it more difficult for manufacturers to create denser drives. This phenomenon is known as track misregistration.

Further, the air gap places some limits on how many platters manufacturers are able to place into hard drives. Hitachi Global Storage Technologies (HGST), an arm of Western Digital, currently uses five platters in hard drives and, with traditional technologies, is releasing a hard drive that boasts 4 TB of capacity... a single unit. That's pretty impressive.

But, it's not enough for the ever-growing demands of exploding cloud providers that need cheap and reliable storage to meet customer demand. To that end, HGST has turned to helium. Rather than forcing air through the drive, helium filled drives will be hermetically sealed to keep the helium in and the bad stuff out. Helium is lighter than air and, as a result, will provide less drag on the platters than air does. Further, because the drives will be fully sealed, they will operate at slightly lower temperatures and in harsher conditions.

With the introduction of helium, HGST indicates that they will be able to add two more platters to helium-based drives. Assuming that the data density stays the same - although HGST indicates that storage density may also increase - that's a 40% increase in platters, meaning that we might see drives approaching 6TB in size next year.

Further, with less drag on the platters comes the ability to operate drives with a little less power. While the power savings from a single drive may not be that much, as you scale this to cloud-provider size, that power savings could be significant. Further, given that HGST estimates that drives may be able to run cooler, it might be possible to adjust data center temperatures up just a little bit, further improving power savings.

Summary

What do you think if this latest trend in increasing hard drive density? Is there something to it or is HGST just full of hot air?

About

Since 1994, Scott Lowe has been providing technology solutions to a variety of organizations. After spending 10 years in multiple CIO roles, Scott is now an independent consultant, blogger, author, owner of The 1610 Group, and a Senior IT Executive w...

64 comments
AribenDavid
AribenDavid

Current drives have filters which allow slow diffusion of ambient air. Sealed drives with dry helium would not have problems with condensation and so could be designed to work at colder temperatures.

AribenDavid
AribenDavid

Helium is very hard to contain, since it is a small molecule and non-reactive. IBM used it in some mainframe logic modules because of its high thermal conductivity.

pgf_666
pgf_666

The earliest work I've been able to find on F/P style cold fusion was by some Swedes in about 1936--their prime purpose was to produce He, not generate power....

danwat1234
danwat1234

Helium is 1 possible way to increase storage capacity. Another is shingled magnetic recording, but this will probably result in slower writes since the whole group of tracks needs to be rewritten once the data on 1 track is updated.


HAMR, perhaps helium is the way to go. Hope none of that helium ever leaks out.

wakanasakai
wakanasakai

@"Why helium". Want some cheese with that wine? Even if I don't dispute the claim that we don't have a lot of helium (& I do), that's a STUPID reason to NOT USE it. That's like me saying that there's only one of me, so I shouldn't be allowed to have a life.

ilovesards
ilovesards

yearh , i agree with sysdev. but even ssd's must be shrunked to make it implantable to our brains. we want the data inside our brains and communicates direct to us without outside interventions. the higgs boson will be a great step . i read a book BRAVE NEW BABY some 30years ago and, yes= we must be star trek transported to some light yaaars distance at light ( minimum) speed ! sure the mechanical drives even if it hits Terra bytes will be rarely useful, if not forgotten .

ubbogus2
ubbogus2

Space and weight are not the issue they used to be. Might as well go back to 5-1/4" drives to expand capacity. But that avoids the issue in that data is still serially read and a mechanical device is searching the areas to be read. Advance the solid state drive and have true 64bit parallel data fed to the CPU.

Darren B - KC
Darren B - KC

Indeed, we're going to see a shortage of helium in the coming years, so this "advancement" (if you can call it that) in the tech is a moot point since it would only make the drives more expensive as helium supplies dwindle. There's also a concern about the availability of rare earth magnets, which are used in hard drives, but that's a whole other topic. Furthermore, as was again pointed out already in another post, why are we focusing on mechanical drives when we should be working on more reliable and inexpensive forms of flash memory for SSD technology?? If manufacuturers insist on keeping platter drives around, I'd rather see more money and research poured into making the drives more reliable rather than increasing the capacities. Seriously, even with my extensive photo and video collection (trust me,... it's a LOT), it still takes me quite a while to fill up a 2TB drive and I'm only just now considering a 3TB drive as my next backup device. But, when I read the cusomer reviews on some of the world's "best rated" hard drives (consumer level, not enterprise class), I'm appalled by the frequency of DOA drives, or drives that completely fail within the first 3 months of use. The failure rate of drives made by companies like Seagate and Western Digital are unacceptably high as they stand right now. Seriously, work on making drives BETTER, not bigger.

rwe9
rwe9

Concerning the comment about using another inert gas, I would say that the atomic weight of the molecules is probably a determining issue, which would make helm the most useful inert gas for the filling of hard drives. Concerning the limited amounts of available helium, I suspect that hydrogen would be also effective in this use, and it is readily available, and at the small amounts in a hard drive, it would not be much of a fire or explosion risk. Concerning the limited amounts of helium, I suspect the amounts to be used in a hard drive case is very small compared to a balloon, so just stop using it in balloons and there will be plenty available. and the market place would much such an allocation, should the availability of helium become an issue.

Dylan Teo
Dylan Teo

If Helium is choosen because it is light then Hydrogen would be a better choice and we have abundant. Fuel cells are already looking at Deuterium which can be extracted from sea water. If not the next gas that we have plenty is Nitrogen.

dAVErSF
dAVErSF

Consider how many hard drives could be manufactured with the helium within one single floating party balloon, sold by the millions daily. That is where the waste is pervasive! Helium is obviously used for at least two basic reasons: it is relatively inexpensive, and being the lowest in density (except for highly flammable hydrogen), it provides the least resistance and narrowest functional gap between the heads and platters

JohnOfStony
JohnOfStony

It's well know that helium molecules are so small that they can pass through glass over time which is why TV cameras using vacuum tubes had a very short life in manned diving equipment which used a helium-oxygen mix to prevent divers getting 'the bends'.So how long before the helium pressure in the drive changes enough to prevent the drives working? I presume this question has been addressed.

sysdev
sysdev

Ladies and gentlemen. You are all probably correct and all probably wasting your time. Technology moves on. Does anyone else remember the first IBM I-GB drive for the mainframe?. That was around 40 years ago and cost $75,000 for 1 GB. Today on my PC (which is a million times more powerful than the mainframes of that era, I have over 10 TB of hard drive storage (pretty inexpensive) and my boot drive is a 1/2 GB SSD. SSD prices have come sown since my machine arrived about a year ago and they are still coming down. I agree with the comment that mechanical moving parts is a problem and that is going the way of the dodo. Just look at your car. Open the hood. If you can see the ground, it is probably an older car. Technology moves on. All the stuff that is now blocking your view of the ground in your car is (I think as far as I can tell) NOT moving parts. I don't worry about it and I view it as a temporary 'idea that works now, - but is not destined to be around forever.partially because of all the reasons mentioned, but also because it is not an issue. IBM has recently publicly announced that they have been able to successfully write and read data at the atomic level just using a very small number of atoms. There are more than enough of those around and the read/write speed is even better than SSDs. I don't know if and when that capability will be commonplace, but remember that your cell phone is probably more capable of processing data than the multi million dollar mainframes of a short time ago and your cell phone cost a whole lot less. This is almost (not quite - but almost) arguing about how many angels can dance on the head of a pin I really think that this is a non-question. Relax. It will be resolved without depleting the currently available Helium (there is a lot more available, just costs mor to get at it) and we will probably (may already have - I am not an atomic scientist) be able to make our own sooner than you think. Technical progress is moving at an increasing rate. Remember the story about the Patent office official who quit in the 1890s because he felt that 'everything that could be invented already had been invented and his job was irrelevant'? Well, he sure was wrong. What we see today will be considered archaic by our own children. People are amazing. Believe it. A couple of months ago, I was startled to watch my 3 1/2 year old granson use an ipad like he was in his teens and had been using it for years. I later found out that his mother had just picked it up that morning from the local library and had received a two minute overview from the librarian (which he had overheard because he was standing next to her) and without a single question, he wasby late morning zipping through all the kid stuff like he had been using it for years. My daughter grew up with computers in the house as both of her parents were and still are in Data Processing and we had PCs and Apple (because that's what the schools had at that time) in the house. My son who is 8 years younger than his sister was using computers before he could read and somewhat before he could really talk well. Stop worrying. This is a non-issue.

DOSlover
DOSlover

If the reduced density of the gas surrounding the drive is the issue, why not hydrogen? It is not an inert gas as Helium is but is is even lighter/less dense and in a hermetically sealed enclosure is not going to cause a fire if it is a pure hydrogen atmosphere. Helium is a bit pricier and I hope includes some process of gas recovery otherwise spinning disk prices will start climbing again to keep step with the capacity increases.

RipVan
RipVan

We will have CHIPMUNKS running the place!

SaintGeorge
SaintGeorge

I just posted a comment about the new technologies not being really new. And then proceeded to read the comments. OMG. Are we a bunch of morons? And by we, I mean YOU. First. Really?? Discussing whether we will deplet the helium in the atmosphere? Making helium from hydrogen explosions? "Need makes the way", really, what next, turning plumb into gold? And second. Solid state, guys!! The problem with hard drives are not air, or dust or whatever. It is MOVING PARTS! All problems developed by hard-drives boil down to broken or degraded moving parts, dead motors, desync between parts. Access times can't go down because of the needs of mechanical access to 0s and 1s. Wobbling and vibrations are unavoidable. This forum reminds me of people arguing how many angels can stand on a pinhead.

SaintGeorge
SaintGeorge

It is not correct that a bunch of tech-geniuses keep saving us from the brink of disaster. Which wouldn't be a disaster anyway, just a slow down. Most, if not all, "new" technologies are not new at all. They have been thought, and researched and tried years and even decades before, but they were too expensive, or required extensive changes to other hardware, or were unreliable, or the technology to mass produce it was not yet available. Usually, it all boils down to costs. Now, whenever we reach an impasse in ANY area of tech, be it hard-drives or sanitary-towels, companies which need to keep growing to post better and better quarters, accept those costs as investment and embrace the "new" technologies.

F4 4ever
F4 4ever

Well, I suppose. But only because we're not tapping our natural gas deposits, which are 7% helium.

sysdev
sysdev

That very well may be correct, but I find it hard to believe that we have a) exhausted the entire supply on our planet (yes - it may be more expensive to find more, but need finds a way) and my guess is that there are other gases that would do almost as well that are available to try. The air we breathe is not clean (even when filtered) and cleaning it better would also help. I understand that it is heavier than helium. Why not Hydrogen? It is even lighter and pretty plentiful. In a closed environment, it is not a fire hazard unless it gets too hot and we know how to lower the temperature sufficiently - we already do it for our CPUs and a closed circuit water cooling system is not that expensive and with more use will be even less expensive.

jon_chalk
jon_chalk

Yes, HE supply is dwindling fast. Once we run out, then what? Though one can argue that very little is needed for the drives, but as demand for the new HE-filled drives, the supply will end. I recommend that the industry use a renewable gas instead.

Ajax4Hire
Ajax4Hire

The best way to make Helium is by combining two Hydrogen atoms (H+H). It is called Fusion and also release lots of energy. Maybe the excess energy from the Helium factory can be used to power the world. Also, He is very small, it will "leak" out no matter how secure and tight the harddrive is. The good news is nothing will replace the He since nothing can leak IN. As for temperature, He will not help, it is surface area and the speed at which you can radiate heat off that will determine the inside temperature. "Simple" but hard to solve mechanical problem.

Ajax4Hire
Ajax4Hire

Helium (HE) is limited, not easy to manufacture. Today is is mined (if you can believe that) out of the ground. The current HE mines are drying? up which is driving the cost of HE up. Most Helium balloons are mostly Nitrogen+Helium, enough HE to make the balloon float.

k9fe
k9fe

Just how often do you have to refill the containment with Helium? Spent over 20 years working with helium and containing it is quite a problem. Unless the containment has been leak checked to 10 to the -10 it is just going to be a slow helium leak. Even at the high level of testing, helium permeates through many materials. Good Luck with this one!

mudpuppy1
mudpuppy1

At least so far. Anyone remember back in the late 1980s? Someone said color laptops will never happen and he was taken seriously. Even I knew that was a pantload. As for SSDs, yes, they will eventually replace platters but as far as I know, they haven't exceeded 1TB with those yet (let alone 4TB) and they are still more expensive. They will get there eventually. Then there is the helium shortage. Also, helium will leak out. I used to work on systems that used helium in cryogenic applications and we used a helium leak detector to make sure the connectors were tight enough to limit the leaks as much as possible.

jf555
jf555

Science fiction, containing Helium is extremely difficult, this will never make it in light of the SDD technology.

swifters
swifters

Errrrr......Have we forgotten about Solid State Drives and the progress they are making. Mechanical is old technology whatever gas you stick in them.

AribenDavid
AribenDavid

@ubbogus2 Because of tolerances and so forth, 5 1/4 would not get you more data stored per drive today. Rotating drives still give much more data per dollar than solid state and will for a long time. They also have no fatigue mechanism. For a test, I placed a solid state drive in swap duty and it failed in less than 2 months. Too many writes

AribenDavid
AribenDavid

@Darren B - KC Good point. Reliability is better now than in the good old days, but we have had drives that are reliable or unreliable for a long time. Early production was in the US and those drives seemed to last. Once production was sent to Malaysia, reliability plummeted. Quality control, humidity, and dust seem to be the factors involved.

ubbogus2
ubbogus2

I wholeheartedly agree. I don't mind having more drives (we can go from C to Z) for capacity. Like the car manufacturer giving a 100,000 mile warranty - give me a hard drive that will be warranted for 10 years.

NickNielsen
NickNielsen

Those are the enterprise-class drives...

NickNielsen
NickNielsen

The reason they are using helium is that it is non-reactive (that's what inert means). Hydrogen is highly reactive and would combine with any compatible molecule, such as any oxygen that happened to leak into the enclosure. I don't know about you, but I sure don't want water inside my drive enclosure!

HAL 9000
HAL 9000

Isn't Inert and is Explosive. Not really the type of thing you want sitting on your Kitchen Table in the Kids NB when it starts to leak. Even a little Hydrogen makes a very big bang as well as having a very low Flash Point. ;) Col

harry
harry

Helium is monatomic and thus has 3 degrees of translational freedom and zero degrees of rotational freedom (hence high "gamma"). This probably makes a considerable difference to its efficacy. The next lightest inert (and monatomic) gas is Neon, with atomic weight about 20, 10 times that of helium.

sysdev
sysdev

Well, we talk back to them of course. That day is already here and not just for us geeks. Every smartphone (well at least the iPhone and Windows Phone) talk to you and listen to you. I have never tried an Android, but my guess is that they have the capability as well. It isn't like the computer on Star Trek, but as Neil Armstrong stated over 40 years ago, 'It is a great leap for mankind' and it is going to get better. Real soon.

HAL 9000
HAL 9000

What type of Angles Christian ones or Weeping Ones? At least the Weeping ones turn to stone when ever anyone looks at them so they are easier to spot and presumably count. Sorry just a little Christmas Humor to lighten the day. :^0 Col

JCitizen
JCitizen

will make using spinning platters obsolete anyway. The storage medium will by more like a clear block of transparent material looking like an ice cube. Lasers will read the electron attitude of each atom in the matrix.

AribenDavid
AribenDavid

@sysdev The government gave up on helium and thought the private sector would take it up. No one has found it economical to separate the helium in natural gas and oil exploration. We will run out if money is not appropriated for helium recovery.

JCitizen
JCitizen

otherwise known as tritium - it makes the process of fusion easier. Just another good reason to mine it from the Moon. Also it would be better to fuse it on the Moon where gravity has less influence, to cause plasma leaks.

DOSlover
DOSlover

Hydrogen in its commonest form has no neutrons. Helium does. Making helium that way is really going to make the drives seriously expensive!

mudpuppy1
mudpuppy1

I had forgotten the value we leak-checked to (been over 20 years). Thanks for the reminder.

gechurch
gechurch

The article is talking about getting the most capacity out of drives. We're many years away from a world where SSDs have the bast capacity per dollar.

JohnOfStony
JohnOfStony

harry@, may I correct your figures? Neon is only 5 times the atomic weight of helium. Helium has a Relative Atomic Mass of 4 (2 protons + 2 neutrons).

HAL 9000
HAL 9000

A computer with He Filled drives speaks how exactly? Col ;)

ubbogus2
ubbogus2

They said that in the mid 80s. Still waiting.

JCitizen
JCitizen

I'd bet those alternatives will permeate the metal even faster. They are very leaky gases.

HAL 9000
HAL 9000

We still have not mastered Holographic Data Storage so maybe some time soon when we give up on spinning platters it may become available. Col