WD Announces 8TB, 10TB Helium Hard Drives

Lucas123 writes: Western Digital's HGST subsidiary today announced it's shipping its first 8TB and the world's first 10TB helium-filled hard drive. The 3.5-in, 10TB drive also marks HGST's first foray into the use of shingled magnetic recording technology, which Seagate began using last year. Unlike standard perpendicular magnetic recording (PMR), where data tracks rest side by side, SMR overlaps the tracks on a platter like shingles on a roof, thereby allowing a higher areal density. Seagate has said SMR technology will allow it to achieve 20TB drives by 2020. That company has yet to use helium, however. HGST said its use of hermetically-sealed helium drives reduces friction among moving drive components and keeps dust out. Both drives use a 7-platter configuration with a 7200 RPM spindle speed. The company said it plans to discontinue its production of air-only drives by 2017, replacing all data center models with helium drives.

Comments

When can we stop selling party balloons

[Anonymous Coward]

Just because congress decided we needed to get rid ot all our heilum supplies quick and dump it for cheap?

Re:When can we stop selling party balloons

[ShanghaiBill]

Just because congress decided we needed to get rid ot all our heilum supplies quick and dump it for cheap?

The amount of helium used in these drives is utterly insignificant. Congress liquidated the helium reserves on the assumption that the market would respond to shortages more effectively than bureaucrats. So far they have been right. Many gas wells that produce helium have be idled, both because of the low price of gas and the low price of helium. So plenty of helium is being "reserved" by leaving it in the ground. If you think you are smarter than the market, feel free to start your own stockpile. When you get rich, you can come back here, post a picture of your yacht, and say "See, I told you so!"

Re:When can we stop selling party balloons

[Electricity Likes Me]

Yes I guess that is why a gas we can only mine tiny quantities of from natural gas reserves is currently being sold off for party balloon use.

Helium is being sold off at any bid price on the assumption that the stockpile is unlimited and needs to be liquidated. It's not being priced according to the long-term cost or capability of supplying the volumes that are being drawn from the stockpile on a sustainable basis, which is what you'd actually do if you were running a business dealing in actual production.

Re:When can we stop selling party balloons

[Rei]

Helium balloons are a minor part of the overall picture. The overwhelming majority of uses are industrial, such as cryogenics. The problem is that they don't recover it. If you want to make a big impact on the helium consumption rate, hard drives is pretty much one of the least effective places you could focus - focus on industrial recovery.

Note that humans will never "run out" of helium. Even if we assume that space-based resource extraction becomes realistic, one can always refrigerate it out of the atmosphere. Or more accurately, refrigerate everything else out and leave the helium behind. There's only a tiny bit in the atmosphere, but for important uses it'll remain a possibility. I saw page that says that neon is $2 per liter. If you're refrigerating neon out of the atmosphere, pretty much all that's left is helium, so you're co-producing it, at a ratio of 3.5 to 1. If we assume that helium demand vastly outpaces neon demand, then the helium cost would be $7 per liter. And maybe less in mass production.

That's not really an absurd price for many uses - such as hard drives. On the other hand, it's dramatically more than today's prices at about $0.005 per liter! You're not going to be making helium blimps at $7 per liter. But if industry learns how to recapture and reuse, they should manage.

(Of course, humans probably wouldn't have to resort to helium extraction from the atmosphere for centuries, pretty much any gas coming out of the ground will be richer in helium than the air)

Re:When can we stop selling party balloons

[ShanghaiBill]

It's not being priced according to the long-term cost or capability of supplying the volumes that are being drawn from the stockpile on a sustainable basis, which is what you'd actually do if you were running a business dealing in actual production.

Yes, clearly all the people investing THEIR OWN MONEY are complete idiots, and you, a random guy on the internet, are so much smarter than actual investors and professional geologists. So does this mean you are going to invest all your money in helium futures, that are obviously going to be worth billions when the helium runs out, just like all the arm chair doomsayers are predicting?

Re:When can we stop selling party balloons

[disambiguated]

Interesting. The last sentence on the Wikipedia page for helium:

"Helium is a finite resource and is one of the few elements with escape velocity, meaning that once released into the atmosphere, it escapes into space."

So how does that work?

Re:When can we stop selling party balloons

[AutodidactLabrat]

So, let me guess. "Government IS the problem" right?
Government created the reserve. Government paid for the refining of He from natural gas in Texas, Oaklahoma and Louisiana.
Government found the program was 1.2 billion in debt.
Government handed the business to Private Enterprise
Private enterprise sold the product dirt cheap (they didn't inherit the outstanding debt, yet another Pro-Capitalist giveaway).
and now, somehow, it is Congress fault?

Re:When can we stop selling party balloons

[Altrag]

By recognizing the difference between "can" and "will." I would hazard to guess that only a tiny fraction of atmospheric helium manages to gain both the required energy and trajectory to actual escape.

Re:When can we stop selling party balloons

[ShanghaiBill]

So how does that work?

The helium in the atmosphere slowly dissipates into space. But it is also replenished by helium leaking out of the ground, where it is generated by radioactive elements emitting alpha particles (which are helium nuclei). At about 5.5 ppm, the source and drain are in equilibrium.

Re:When can we stop selling party balloons

[wonkey_monkey]

Yes I guess that is why a gas we can only mine tiny quantities of from natural gas reserves is currently being sold off for party balloon use.

Party balloon helium - or so the argument goes - is an impure by-product of the production of industrial-use helium.

Re:When can we stop selling party balloons

[disambiguated]

I suppose it's going to be a while before we run out of alpha emitters. So the Wikipedia page is wrong then, when it says Helium is a finite resource. Last time I trust Wikipedia (yeah right:).

You said it slowly dissipates into space. That means the rate it leaves the atmosphere is low, so the rate it is replenished is low, and that's the limiting extraction rate.

According to this (that didn't take long), the rate Helium leaves the atmosphere is 50g/s, or 3e5 cm^3/s. The National Helium Reserve is 1e9 m^3. So, extracting all of the Helium from the atmosphere before it escapes, it would take 1e9 m^3 / (3e5 cm^3/s), or over 100 years to replace the reserves.

But extracting all of it is hopelessly unrealistic. I don't know, but it seems even 1% would be ambitious. So now we're looking at tens of thousands of years.

So either the national reserve is ridiculously large, or removing it from the atmosphere is not going to be a solution to the shortage. Right? Or am I missing something (else)?

Re:When can we stop selling party balloons

[stevelinton]

Interesting. The last sentence on the Wikipedia page for helium:

"Helium is a finite resource and is one of the few elements with escape velocity, meaning that once released into the atmosphere, it escapes into space."

So how does that work?

At a given temperature the typical velocity of a gas molecule depends on its mass. The lighter, the faster.
Helium is the only gas molecule that is stable in the atmosphere and has a typical velocity near the top of the atmosphere that is faster than Earth's escape velocity, so it slowly diffuses up to the top and then is gradually lost to space.

Re:When can we stop selling party balloons

[Enry]

Is that different from weapons-grade helium?

Re:When can we stop selling party balloons

[Jason Levine]

Part balloon helium raises your voice.

A weapons grade helium bomb raises the voices of everyone in a five mile area to the point that it's above the human hearing range, thus disabling the enemy's communications.

Obviously, we need to watch out for potential terrorists in the form of people holding balloons.

Don't even get me started on the weapons-grade ice cream and the disabling headaches is causes.

Re:When can we stop selling party balloons

[Rei]

Helium exists in the atmosphere not because of the helium reserve, but because the planet constantly outgasses it. It's a product of the radioactive decay chains within the planet.

And if it costs $7 a liter, you better believe people will consume it a *lot* slower. Mainly recapture, but also less frivolous usage.

Re:When can we stop selling party balloons

[Electricity Likes Me]

It's not being priced according to the long-term cost or capability of supplying the volumes that are being drawn from the stockpile on a sustainable basis, which is what you'd actually do if you were running a business dealing in actual production.

Yes, clearly all the people investing THEIR OWN MONEY are complete idiots, and you, a random guy on the internet, are so much smarter than actual investors and professional geologists. So does this mean you are going to invest all your money in helium futures, that are obviously going to be worth billions when the helium runs out, just like all the arm chair doomsayers are predicting?

Will you take the futures option the other way? That the current prices are going to stay at the level they are for say, the next ~20 years plus or minus 10%?

I don't know who you think is investing in this. People are just buying helium because they need it. The smart people are investing in gas mining because that's always profitable, and the big companies building huge helium liquefaction facilities are absolutely betting the price is going to rise in the near future.

Apparently the free market thinks you're kind of an idiot about this.

Helium?

[bobstreo]

I thought we were running out of Helium reserves?

Are there not other available inert gases that work as well?

Re:Helium?

[Anonymous Coward]

Certainly seems like a technological dead end.

Re:Helium?

[nitehawk214]

Considering a single balloon has more helium than $10,000 of these hard drives, I don't think this is going to be a serious issue anytime soon.

For party balloons, well, too bad.

Re:Helium?

[Gestahl]

I was also wondering this. Wouldn't nitrogen or argon/neon be cheaper?

Re:Helium?

[Anonymous Coward]

No, there are not. In this case, low molecular weight is the key, and that nearly rules out anything except H2 and He.
H2 is too reactive though. Ne is interesting, but even more expensive the He and the only advantage is less leaking.
N2 is not meaningfully different from normal air, and Ar is even heavier.
CH4 is cheap, light, and mostly unreactive (at moderate temps) but it's really not light enough to compete with He.

And that's the end of the list.

Re:Helium?

[confused one]

Helium is the second most common element in the universe. OK, there's not a lot on Earth... but it's easy to make.

Re:Helium?

[fuzzyfuzzyfungus]

That would be interesting to know. Helium is remarkably low-density by the standards of inert gasses(~.18grams/Liter vs. ~.9 for Neon and 1.25 for the ignoble-but-pretty-well-behaved Nitrogen), so it may actually be substantially different from any of the other options.

Re:Helium?

[Ralph Wiggam]

You have a weird definition of "easy".

Or are you talking about extracting it from natural gas wells?

Re:Helium?

[MavEtJu]

Now if there was an easy way to get it out of all these stars....

Re:Helium?

[Anonymous Coward]

Nope, must be low density and viscosity. Only hydrogen is lighter, and that has problems...
Or vacuum, which I wonder why not? Maybe the lubrication would fail in vacuum?

N2 or Ar are heavier and denser than He.

Re: Helium?

[pollarda]

I usually wrap one in a towel and then twist until it pukes all its helium -- kinda like juicing an orange but bigger.

Re:Helium?

[AchilleTalon]

Easy? Do you know how helium is produced? The helium we currently have access to is the result of Earth crust radioactive elements disintegration.

How do you think you gonna produce efficiently and in a cost effective manner He in volume? Tell us, I am very interested to start a business to produce it and I will give you half the shares for your effort.

Re:Helium?

[cheater512]

It has to do with the actual atom sizes, not the fact the gas is inert.
Their goal is for lower friction which Helium provides.

Now Hydrogen could be quite interesting as a replacement..... :D

Re:Helium?

[Anonymous Coward]

But helium filled disks will ensure an early end of life though. As there is basically nothing you can build that can contain it indefinetly..

Re:Helium?

[Ikester8]

I always thought that a pile of alpha-emitters could serve as a source of helium, since alpha particles are helium nuclei, but it would take a prohibitively long time to generate much.

Re:Helium?

[Bengie]

Vacuums don't conduct heat very well.

Re:Helium?

[BradMajors]

Hot air works.

Re:Helium?

[BradMajors]

Speak with the Qatarians:

"Air Liquide recently started up the world’s largest helium purification and liquefaction unit, a turnkey project at Ras Laffan Industrial City, Qatar. The new unit’s production capacity is approximately 38 million cubic meters of helium per year."

http://www.airliquide.com/en/q...

Do I get half the shares in your business?

Re:Helium?

[ShanghaiBill]

I was also wondering this. Wouldn't nitrogen or argon/neon be cheaper?

The point of using helium is NOT that it is inert, but that it is low density (although the inertness is also nice). Neon is five times as dense and far more expensive. Methane is four times the density. The only gas lighter would be hydrogen. But hydrogen has a nasty habit of migrating through metal, leaking out, and embrittling the metal in the process. Low density gases reduce friction both through reduced mass, and a higher speed threshold for laminar (rather than turbulent) flow. Low density gases tend to also be better heat conductors, helping to keep the disk cool. That is why high density gases, like xenon or sulfer hexafloride, are used in insulated windows.

Re:Helium?

[fnj]

How the hell does this address GP's point that all the helium produced comes from subterranean gas? Way to miss the point.

Re:Helium?

[fnj]

Like maybe all the alpha emitters ever created might produce enough helium in a billion years to fill a few party balloons?

Re:Helium?

[Anonymous Coward]

But helium filled disks will ensure an early end of life though. As there is basically nothing you can build that can contain it indefinetly..

And who the hell plans on running a data center hard drive indefinitely?

Once your users do their best imitation of Hoarders on the new file server, you'll be forced to replace those "tiny" 4TB drives 3 years from now with 12TB ones because they can't ever seem to find the fucking delete button, even if they had three YouTube instructional videos and a gun to their head.

End of life? Yeah, due to filling the damn thing, not wearing it out.

Re:Helium?

[Mr. Freeman]

No surprise there. WD doesn't want people not re-buying their hard drives every few years.

Re:Helium?

[Anonymous Coward]

How does the helium not leak through the seals over time? It's a very small molecule. Maybe they weld the cases shut.

Also, if that problem is solved, why not use hydrogen? It is flammable , but only if it leaks out, and it's only a small amount.

Re:Helium?

[FatdogHaiku]

Maybe the lubrication would fail in vacuum?

Well, that would suck...

Re:Helium?

[0123456]

Uh, no. Helium is one of the most common byproducts of nuclear decay. We just call it 'alpha particles'.

Re:Helium?

[Z00L00K]

Well - considering the amount of helium literally wasted in balloons this is at least a decent use. But otherwise I agree. Next step would be hydrogen drives.

Re:Helium?

[Man On Pink Corner]

I'm also curious how they keep the helium locked up in the drive housing.

Hydrogen is more reactive, and causes some metals to become brittle. Not sure if either of these would be a problem in hard drive applications, though. Presumably the manufacturers have done their homework.

Re: Helium?

[Anonymous Coward]

I think you're thinking of Seagate there

Re:Helium?

[AutodidactLabrat]

It's easier to use titanium than carbon composites, titanium being one of the worst hydride offenders. Iron, of course, suffers embrittlement in H2 as well. Neither of these problems is insurmountable.

Re:Helium?

[SuricouRaven]

It's been done. That's how it was proven that alpha particles are helium nuclei. Someone put an alpha emitter in a very well-sealed vacuum container, waited a long time, and was able to detect a trace of helium gas that couldn't have come from anywhere else.

Re:Helium?

[Altrag]

Most likely they don't. They just have to keep it escaping slow enough that the expected lifespan of the drive runs out before the helium does. That's only a few years.

Re:Helium?

[Altrag]

Nothing else in the drive lasts indefinitely either, so as long as they keep the helium dense enough to outlast the other components in the drive, this is a non-issue.

Re:Helium?

[FireFury03]

And who the hell plans on running a data center hard drive indefinitely?

ISTR that the big datacentres, such as Google, run drives until they fail - the systems are redundent enough to cope with a failure with no problems and they have so many drives that it's more cost effective to have a resilliant system and just run the drives into the ground than it is to preemptively retire them (and still have to cope with unexpected premature failures).

Re:Helium?

[Anonymous Coward]

That's why the Republicans in 1996 ordered the US to destroy its helium reserves by 2015

If you had actually read the page you link to, you would have known that the U.S. government did not order to destroy, but to sell the helium reserves, because they rightly considered that selling gasses for industrial use is not a government task.

Re:Helium?

[Mal-2]

Methane is still a large improvement over air or nitrogen, however.

Air is, to a rough approximation, 80% diatomic nitrogen and 20% diatomic oxygen -- or molecular masses of 28 and 32, respectively. The average molecular mass of air is then about 28.8. (You can quibble about the numbers but this is close enough to make the point.)

Methane -- CH4 -- has a molecular mass of 16. While it's quite a bit higher than helium (molecular mass of 4), it is still a lifting gas. In fact, one night over drinks, an engineer friend and I decided to do some back-of-the-napkin calculations based on an absurd idea I came up with. I wanted to know if it would be possible to lift a cow by trapping its own emitted methane, and if so, how long it would take. It turned out our answer was somewhere on the order of ten years.

Ammonia -- NH3 -- would weigh in at a molecular mass of 17. Unfortunately, it tends to be horribly corrosive in a lot of situations.

Re:Helium?

[Charliemopps]

Give me a break.
411 voted Yea, including 180 democrats.
Only 10 democrats voted Nae.
https://www.govtrack.us/congre...

This is the very definition of a bi-partisan bill. Your attempt to demonize the republican party and therefor cast your party, the democrats, as some kind of protector of the people is a joke. The democrats are just as corrupt as the republicans. And... this wasn't a bad bill anyway! The US government shouldn't have control over the entire supply of an Element!

Re:Helium?

[impossiblefork]

It would probably work quite well, but it might impose additional constraints on how the hard drives could be built since hydrogen can enter metals. There's of course no risk of explosions (after all, these things exist: http://en.wikipedia.org/wiki/H...), but it could certainly be interesting.

Re:Helium?

[hairyfeet]

Does it come with some sort of gauge, so you can know when its getting close to running out? Because I don't want a drive with a time limit where I don't have a way to keep up with the countdown.

I'm sure I'll hear a bunch of analogies but I've dealt with more drives in the shop than most guys have had hot meals and with modern drives? Unless you drop the thing you usually get plenty of warning before failure, not to mention that having to get 8TB worth of data off while the drive dies does NOT sound fun so having plenty of warning would be good. Heck I've had 1TB drives in the shop that was a race to get the data off before they croaked, 8-10TB sounds like a nightmare.

Re:Helium?

[Eunuchswear]

A single MRI machine needs 10,000 liters of helium.

So it's a good thing Air Liquide are setting up a plant to extract 3.8E10 litres a year, enough for 3.8 million MRI machines.

Looks like the Republican health care destruction plan has failed.

Re:Helium?

[Bob the Super Hamste]

I doubt this is actually true. This may be the official line from management but having several relatives who work in hospitals one of whom works in the maintenance area it seems this should be treated similarly to all other statements from the piss poor managers at hospitals, aka 100% Grade A USDA certified BS. If this were true you are telling me that the hospital is willing to risk damaging a machine that is in the 1/2 to 2 million dollar range by not doing maintenance to try and save maybe a couple of hundred dollars on helium. Mangers have tried to create similar situations at the hospital my father worked but once they get up to higher levels quickly get sorted out since the low level management likes to fight over their little fiefdoms and bonuses. This seems to stem from the fact that the biomedical technicians managers don't understand that the biomeds were a cost center and would try to cut costs so that the department would show a profit.

Of the more notable instances one was when the manager tried to cut back on the amount of hose for they were buying for the anesthesia machines since they were buying a fairly substantial amount of it. The problem is that it has to be replaced periodically and they know how many anesthesia machines the hospital has so they were basically replacing what they used and keeping a fairly small reserve. Then there was the pissing contest another manager of the biomeds got into with a manager of the surgery department over if some expensive device should be repaired (cost comes out of the biomeds department) or replaced (cost comes out of the surgery department). To repair the device was something like $1000 in parts but replacing it was over $70,000 and this got escalated up to the directors over the course of a couple of weeks.

I have a feeling that your situation will turn out much the same way once low level managers get slapped around a bit by upper managers who's fiefdoms include the whole hospital and would rather not buy and new MRI machine because maintenance was not done on their already existing and functional one.

Re:Helium?

[Anonymous Coward]

What about the Democrats who could have countermanded the Republican order? The Repbulicans haven't held power all the time since 1996.

Re:Helium?

[countach]

It's counter-intuitive that a low-density substance would conduct heat well. It seems the opposite with solids, where low density substances are good insulators.

Re:Helium?

[Eunuchswear]

What brain dead fuck moded that "troll"?

Re:Helium?

[K. S. Kyosuke]

Helium don't usually actually conduct the heat (unless somehow fixated), it convects the heat. And the reason why helium convects heat well is that it has a low molar mass. A high-thermal-capacity gas with fast molecules being good at convecting heat doesn't seem counter-intuitive to me.

Re:Helium?

[K. S. Kyosuke]

Saturn would be easier, wouldn't it?

Re:Helium?

[ChrisMaple]

The low density "solids" that are good thermal insulators are (generally) not really solid: they're mostly air kept from moving by a mess of tiny strands of glass or wood particles, or bubbles containing gas or vacuum.

Re:Helium?

[ChrisMaple]

The youtube videos have to be ultra-high definition, stored in a lossless format at 1000 fps. Copies in all known languages including Klingon. Each must be saved on the drive in triplicate and also stored on a backup drive. Oops! need another drive.

Re:Helium?

[ChrisMaple]

Bad sector remapping extends life by allowing some sectors to fail without the drive failing.. Crippling remapping would cause an increased return rate.

Re:Helium?

[nitehawk214]

Think about it. The volume of 12 hard drives would easily fit instead of a standard party balloon. Now consider that most of the drive is filled up with platters and electronics, and there might be 10 cents worth of helium inside of each hard drive. Lets say they use super pure helium... so 50 cents.

Re:Helium?

[Eunuchswear]

Why?

Re:Helium?

[Neil Boekend]

1. Generate all electricity from nuclear sources.
2. Store the alpha emitting waste in a capturing facility.
3. Wait for the alpha particles to encounter electrons (shouldn't take long)
4. Siphon the generated helium off.
5. Purify sufficiently.
6. ...
7. Profit.

Re:Helium?

[Luckyo]

Essentially all drives already have many such gauges. Just take a look at SMART readings of any modern hard drive. They have several "pre failure" and "old age" counters.

This would simply have to measure temperature, which would increase with increasing friction from loss of helium, just like all modern drives do, and start giving out SMART warnings after certain threshold is crossed, just like modern drives already do.

Re:Helium?

[Luckyo]

It addresses his claim of supposed lack of cost-effectiveness.

Re:Helium?

[confused one]

Of course I know where it comes from... It's extracted from natural gas as a byproduct and initially helium comes from alpha decay of radioactives, as you say. Producing it through radioactive decay, that's the slow way. Helium is produces in a number of fusion reactions. They don't have to be energetically favorable to produce helium. Now... if you want cost effective, then that's another problem.

Re:Helium?

[hairyfeet]

If you trust SMART I have a bridge you may be interested in as I have dealt with drives at the shop where you could literally hear the head slamming that SMART said was just hunky dory. You also can't trust any kind of "old age" metric because that ignores the simple fact that some companies build better products, it would be like judging the reliability of all cars based on what the bottom of the line Kia is like in 5 years. Hell when it comes to reliable I will take a used or refurb Samsung over a brand new Seagate consumer line drive as I've found everything above 500GB on the consumer line is buggy firmware and flimsy as hell, I swear you can look at a Seagate 1.5Tb funny and it'll commit suicide.

At the end of the day surely to God it can't be THAT hard to put in a simple helium gauge so you would know its going down and by how much, in fact the only reason I could see not putting one in is to fuck folks over on the warranty.

Re:Helium?

[Luckyo]

I do trust SMART for old age/wear and tear problems. I also have 2x100GB seagate 7200.7s from 2000 that are still running, as well as several 1GB 7200.12s.

They're been through a lot more than funny looks, like having my entire PC tower fall on its side accidentally while they were working and other similar things. They still work fine.

SMART is generally designed to track certain kinds of issues. Quick breakdown of controller logic isn't one of them, and state of bearings on the write head isn't one either. But wear and tear aspects of it, like temperature increase from helium escaping the drive? Absolutely.

Data Distortion

[Anonymous Coward]

All my audio files sound like "Tiny Tim"

USA has it.

[turkeydance]

helium, that is. invisible gold, Texas A (for Amarillo).

Re:USA has it.

[AutodidactLabrat]

Yep, and we're running short. Soon, all research into superconductivity will have to stop. Have to wonder why they aren't use H2. A little heavier, but not much and as long as you don't use Iron, titanium, vanadium or similar metals which form hydrides.....
Has to be one of the ultralight gases though. Speed of sound and all that.

Re:USA has it.

[fnj]

"One of the ultralight gases" - well, that narrows it down to H2 and He and nothing else, now, doesn't it? You left out that the candidate gas has to be very high in thermal conductivity as well, but both of those are so.

BTW, hydrogen is not higher density than helium. It is considerably lower. It would be more ideal than helium for this use if it weren't for its highly undesirable propensity to react with other substances.

We're not "running short" of helium. The ready supply of helium-rich natural gas is going to run out some time in the future, but so is the ready supply of just about everything else.

All the helium you could ever possibly want is in the atmosphere. 0.0005% by volume, or 0.00007% by mass, of the atmosphere is helium. The total mass of the atmosphere is 5x10^18 kg, so the total mass of the helium in the atmosphere is 3.5x10^12 kg. At STP (standard temperature and pressure), that represents 2x10^13 cubic meters.

That would fill 100 million Hindenburgs, or many trillion party balloons.

So no, even if/when the current sources of helium run out entirely, "all research into superconductivity" will NOT have to stop. It will "just" cost more.
The helium in the atmosphere is constantly escaping into space, and constantly being refreshed by escaping from the earth into the atmosphere. Inside the earth it is constantly being produced by nuclear processes.

Yes, it would be far more expensive to extract helium from the atmosphere at 0.0005% concentration than it is from natural gas sources at concentrations of from 0.1% to several %. But it is glaringly obvious that it is not impossible. The concentration of neon in the atmosphere is less than four times that of helium, and ALL the neon produced is produced by extracting it from the atmosphere.

Re:USA has it.

[AutodidactLabrat]

There is no single H at stp. It is always H2 and the molecule, while weighing less than an He 'molecule', is larger than He atom, resulting in lower thermal transfer.
THAT being said, single H does occur in elevated heat or electric field potential, thus invading grain boundries in iron alloys, and causing embrittlement thereof.

Re:USA has it.

[dave420]

The helium is constantly leaking into space and being replenished by subterranean sources. When those sources are gone, so will the helium in the atmosphere. Your 0.0005% figure will decrease over time.

Re:USA has it.

[fnj]

Yeah, come back to me in several billion years and we can commiserate over it. An utterly pointless concern over relevant time periods. Because after several billion years, there either won't be any human race, or it's capabilities will be beyond our comprehension. Those people will just mentally command a matter synthesizer to make helium, or they will grab it from interstellar space. There is essentially an infinite quantity of helium in space, and it isn't going anywhere until the universe ends.

Re:USA has it.

[fnj]

Hydrogen thermal conductivity = 0.1805 Wm1K1
Helium thermal conductivity = 0.1513 Wm1K1

Now, which one is higher again?

Re:USA has it.

[Neil Boekend]

Those subterranean sources are being replenished by natural alpha emitting sources.
When those sources are gone the sun will have gone red giant, absorbing the earth. We should have enough helium by then.

Re:USA has it.

[AutodidactLabrat]

I stand corrected

containment

[Eric Coleman]

The most important question is what is the lifespan of the helium containment. Helium is notorious for getting in to and out of places that other elements can't. For example, in balloon borne cosmic ray experiments, or anything with a calorimeter or hodoscope that utilizes photomultiplier tubes, you have the problem of the helium from the balloon getting into the PMTs, which hold a vacuum. Of course, there are low pressure conditions to consider, but I'm still skeptical of the helium staying in the hard drive.

Re:containment

[anotheryak]

Totally agreed, my first thought was "how will you keep it filled"?

Re:containment

[ThatsMyNick]

It is probably refillable (or I hope it is). Like every year you get to refill your helium. Containment is always a problem with helium, usually solved by regular refilling.

Re:containment

[TechyImmigrant]

Not a problem. The MTTF of a 10TB magnetic disk is probably around 10 minutes.

Re:containment

[Anonymous Coward]

This takes "planned obsolescence" to a whole new level. "Your helium levels will reach critical in 32 days. Please replace your drive with a new one now."

Re:containment

[maswan]

Clever materials choices and lower pressure than on the outside (~40% IIRC). Luckily leakage is easily measured in the product design and testing phase, as well as ongoing QA. So not nearly as much risk to your data as stupid firmware bugs that only turn up under some circumstances after lots of usage. And no, they won't be refillable.

Re:containment

[Anonymous Coward]

The enterprise grade drives are designed to be refillable. Consumer grade will not be, but will have a several year MTBF by which time the drive will be obsolete anyway and you'll be able to get one 4X the size for the same price.

So not really an issue.

Re:containment

[confused one]

Just fill the datacenter with helium... That way as much will leak into the drives as leaks out.

Re:containment

[fuzzyfuzzyfungus]

You just buy a WD-authorized helium cartridge... Make sure that it hasn't dried out, and that the activation chip hasn't expired.

Re:containment

[sexconker]

The enterprise grade drives are designed to be refillable. Consumer grade will not be, but will have a several year MTBF by which time the drive will be obsolete anyway and you'll be able to get one 4X the size for the same price.

So not really an issue.

Yeah, because I'm going to have keep a helium tank on hand to refill my drives / call a WD tech to do it for me with hs own supply of helium - "Go ahead and open up the drive bays in our secure data center and jam your nozzle into our running drives, you said you were coming between 8 AM and 5 PM, right?".

Re:containment

[ColdWetDog]

What's wrong with a party balloon and a hose?

Re:containment

[aliquis]

I guess it's a better usage than freaking balloons but should we care about that way are using it?

Or just assume we'll go all in on fusion later?

Re:containment

[Anonymous Coward]

> Clever materials choices and lower pressure than on the outside (~40% IIRC).

Why would lower pressure make a difference? IIRC that is not how partial pressure works.

Re:containment

[ebonum]

Simple question: 10-15 years from now, after the helium has escaped, will this thing still work? Today you can take a HD built 10-15 years ago, and there is a good chance it will still work (assuming you have the right cables, drivers, etc.).

Re:containment

[rtaylor]

That's almost exactly how it works with our quarter million dollar SANs.

They get paid to maintain the SAN and regularly visit to swap hardware bits or apply software patches to it.

Re:containment

[maswan]

Yeah, not so much for permeation, maybe, but they still quote this as one of the things that help. Could maybe do some for actual leaks - but getting air in would be sad too. The big reason for lower pressure is the lower resistance though - I like paying less in power&cooling thanks to lower power use to keep the platters spinning.

Re:containment

[ShanghaiBill]

The enterprise grade drives are designed to be refillable.

I am very skeptical. TFA does not say anything about refillable drives. If there is a valve for refilling, then that valve itself will be a major point of failure for helium leaks. It is very difficult to design a valve that will let helium flow in, without the helium also leaking out.

Re:containment

[x0ra]

Will SATA, or whatever connectivity still be around 10-15 years from now ?

Re:containment

[geekmux]

This takes "planned obsolescence" to a whole new level. "Your helium levels will reach critical in 32 days. Please replace your drive with a new one now."

I'd say this more takes IT cost justifications to a whole new level...

"Uh yeah hey boss, we're gonna need to buy $8K worth of replacement hard drives this month."

"Wait, what? We didn't budget for that! Why?!?"

"Well, I just got an alert, ours are leaking pretty bad. Low PSI warning. These things aren't SSD run-flats you know."

I mean c'mon...that would peg the bullshit meter of even the PHB.

Re:containment

[fnj]

Yes, sealing of the valve is a non-trivial issue, but it clearly can't possibly be an insuperable problem. Every K tank of compressed helium gas produced since the 1920s has such a valve, and those tanks are pressurized to well over 100 atmospheres. I have had such a tank sitting for over 10 years with negligible leakage as measured by gauge pressure. I did learn to my dismay that if you leave the main valve open and rely on the regulator and balloon-blowing attachment to hold, you will wonder where the gas went within weeks to months.

It would be interesting to know the pressure these drives operate at. If it is just room pressure, then I don't see how you could refill it unless you had an outgress valve as well as an ingress valve, in order to flush it.

Re:containment

[Anonymous Coward]

Not enough DRM.

Re:containment

[TheGratefulNet]

obsolete?

if you use drives as shelf-spares or backups, then this is a MAJOR problem!

I have drives that are 10+ yrs old and while I don't spin them up very often, I do expect them to still work years from now as long as I give them a spin-up every so often, to keep them in shape.

a drive that fails just sitting there, unused, is NOT something I want to buy! or own.

Re:containment

[timeOday]

Perhaps the drive can still run on air, albeit with more power consumption. In googling it, the advertised benefit is reduced turbulence and drag.

Anyways, the hard drive is rigid (unlike a balloon) so for helium to escape, something must leak in or pressure will drop. I can't imagine all the helium would leak out creating a total vacuum.

Re:containment

[wooferhound]

Are these drives pressurized with helium or at normal atmospheric pressure ?

Re:containment

[ShanghaiBill]

It would be interesting to know the pressure these drives operate at. If it is just room pressure, then I don't see how you could refill it unless you had an outgress valve as well as an ingress valve, in order to flush it.

WD's website says the helium is "low pressure", whatever that means. But regardless of the pressure, when the helium leaks out, it will not be displaced by air. It will leave behind a vacuum. The helium will leak out, but nothing will leak in to replace it. It will stop leaking when the partial pressure equals the concentration of helium in the atmosphere, about 5.2 ppm.

Re:containment

[SuricouRaven]

Can anyone run calculations on this? I imagine it'll be welded shut, so you're looking at the helium having to escape through perhaps a 5mm-thick block of the most impermiable metal they can find. It may well be that it does leak, but only after long after the drive would be considered obsolete. You might not want to use them for archival storage, but how many production drives will still be in use after twenty years?

Re:containment

[dissy]

Why not just fill the whole data center with helium... that way, if any gets out of the drive, it's quickly refilled from the outside :-)

Yeah, that will work.... until the data center floats away.

Sorry boss, as much as I've tried to fight the fad, our data center is now in the cloud...

Re:containment

[Ungrounded Lightning]

But regardless of the pressure, when the helium leaks out, it will not be displaced by air. It will leave behind a vacuum. The helium will leak out, but nothing will leak in to replace it.

(Except maybe hydrogen, but there's not much of that in your local air.)

So your metal parts vacuum-weld and tear themselves apart starting at the contacting surfaces, and adding lots of hydrogen to the air around the drives just makes the parts become brittle on their way to failure.

Re:containment

[K. S. Kyosuke]

But for it to become a cloud, it's not enough to just float in the sky, it also has to contain a lot of vaporware.

Re:containment

[Neil Boekend]

When the helium is gone the drive will be high vacuum. The drive should include a breakaway disk (with shrapnel filter) to prevent collapsing of the enclosure. The system should detect that break and slow down the drive to 5400 RPM to fail gracefully.
SMART should report the pressure inside the drive and the status of the breakaway disk.

Then you can get the data off while the drive is in degraded mode.

Re:containment

[Neil Boekend]

Still, that is approximately what happens.
Partial pressure put simply:
Forget absolute pressure. What matters in leaks is the pressure of each gas individually.
There is 0.00052% of He in the atmosphere. Assuming 1 bar of absolute pressure this means there is a partial pressure of 0.0000052 bar of helium in the atmosphere.
If the drive is filled with 1 bar absolute pure helium the difference will be 1-0.0000052 = 0.9999948 bar. That is the pressure that is important. There is no way the helium will not leak out. There is no such thing as a closed system. The system will also leak air in but far less. The result is a vacuum in the drive. In the end that vacuum will be filled with air but that takes far slower.

Now how long will the loss of helium take?

A 3,5 inch drive is about 0.3l. I assume that half of that is filled with hardware so I assume 0.15l He. A properly welded system without any connectors is probably in the range of 10^-12 mbar*l/s He leak tight. If I assume the drive will work at 0.5 bar helium we can take 10^12*500 mbar *0.3l = 150 * 10^12 seconds.
A year has 3.15 * 10^6 seconds. That is almost 5 million years. Not really something to worry about.
Yes there is a leak. Yes the helium will escape. No it doesn't matter because it just takes extremely long.

With a bad weld the time would drop significantly. However, the detection is easy. Helium leak detectors are commonplace, to detect minute leaks in high purity systems.

Sources:
Dimensions drive: Wikipedia.
Leak rate: I can get connectors to 10^-11 as standard items (Swagelok VCR full metal seal). Fully welded systems are probably better than that.

Re:containment

[Neil Boekend]

I really should have done the calculations before posting that. It takes almost 5 million years to lose half of the helium, assuming a proper weld.

Helium loss is no issue.

Calculations are in my reply to this post

Re:containment

[gizmod]

Ok this might be a really stupid highly uninformed question, but how long would a perfectly spherical graphene balloon contain helium at X PSI? Perhaps 2 layers one slightly offset to the other one to cover additional gaps?

well of course they did!

[sribe]

I ordered 6TB drives 3 hours ago...

Re:well of course they did!

[Anonymous Coward]

If it has only been 3 hours, then you can still cancel your order. :P

p.s. Multiple-of-3 / half-generation HDDs have terrible service records. (Hint: They basically just add a 3rd platter to a 2-platter design.) That's another reason to cancel your order. ;)

Re: well of course they did!

[alen]

Not with amazon you can't
Preparing for shipping as soon as you order and it wont ship for another day

Re:well of course they did!

[fnj]

p.s. Multiple-of-3 / half-generation HDDs have terrible service records

That generalization is highly questionable. I have had a population of over twenty Toshiba 3 TB, including 12 spinning 24x7 that have from 6-14 months of power-on-time. Zero in-service failures.

Re: well of course they did!

[Richard_at_work]

Odd, I cancelled an Amazon order yesterday, 6 hours after I placed it with Prime delivery. Cancelled just fine, even tho it was in "preparing for shipping", they just don't guarantee that it will actually be cancelled until it is.

Re:well of course they did!

[ssam]

High capacity 3.5" disks use 5 or 6 platters.

Re: well of course they did!

[ssam]

If you are in the UK you are covered by distance selling laws, so you can return an un-opened product within 14 days.

Re:well of course they did!

[toddestan]

Well, later on you get the cut-down version, where the 3TB is (for example) a 2-platter 4TB drive where one side of one of the platters is not used for whatever reason.

With that said, I have four of the Seagate 1.5TB drives that seemingly have earned quite a reputation. but mine thus far have been flawless.

Only one problem...

[Chris Mattern]

Now all my MP3s sound like the Chipmunks!

Leakage?

[seven of five]

Helium, being not only a small atom but a monatomic gas, leaks through the tiniest holes. What happens when the helium gets out?

Re:Leakage?

[BringsApples]

If it does leak out, the drive will surely crash, and then you lose your data, and go buy another drive(s). Seriously though, this is compacting more hardware into the same space as current technology, in order to fit into the current frame of things as they are, which is nice. But in my way of thinking, this is a lot more "stuff" crammed into a space that should maybe be bigger, and would seem to make crashes more likely. I bought a 2TB drive not long after they came out, had it in a RAID setup, and both drives crashed 3 days after I installed them.

I'm one that normally complains about (everything!) changes to hardware, but I wouldn't mind having larger hard drives, if it means having more stability.

Re:Leakage?

[BringsApples]

I bought a 2TB drive not long after they came out, had it in a RAID setup, and both drives crashed 3 days after I installed them.

I mean I bought (2) 2 TB drives.

Re:Leakage?

[NJRoadfan]

Remember the Quantum Bigfoot? It was big, but it was cheaper than the competition since it was 5.25" form factor. They turned out to be pretty reliable too.

Re:Leakage?

[demonlapin]

It was hellaciously slow, though.

Re:Leakage?

[pjwhite]

I still have a couple of Bigfoot drives running strong in a Windows 98 machine. Most reliable drives I've ever had. Slow, yes, but for my application, fast enough.

What about heat-assisted magnetic recording?

[amaurea]

When these drives were first announced it was speculated that they would use heat-assisted magnetic recording, which could store a bit into a single magnetic grain rather than a domain consisting of hundreds of them. But it turned out that they used shingled magnetic recording instead, which seems to have less long-term promise. What's the news on HAMR? Is it still being pursued?

Re:What about heat-assisted magnetic recording?

[Anonymous Coward]

I tried using HAMR, but my drive is just in pieces now.

Re:What about heat-assisted magnetic recording?

[jeffb (2.718)]

What's the news on HAMR? Is it still being pursued?

Don't stop. It's not HAMR time.

Data recovery?

[Anonymous Coward]

How much harder and more expensive will the helium make data recovery?

Re:Data recovery?

[arbiter1]

none, they just replace all the air inside the drive with helium, they can open the drive up without a problem.

Re: Data recovery?

[Anonymous Coward]

Remember when you were a child and your balloon floated away, never to return?

Just like that.

Re: Data recovery?

[Anonymous Coward]

Remember when you were a child and your balloon floated away, never to return?

Your data will go just like that.

SSDs will outpace platter drives

[BenJeremy]

By 2020, SSDs will have greater capacities than 20TB.

We are seeing the buggy whip manufacturers in full denial. 10TB drives should have been out a year ago, and consumer 6TB drives should be selling for under $100. The floods in Thailand gave platter drive makers an excuse to keep the prices (and profits) jacked up artificially while the insurance money replaced aging plants with the latest technology.

With a fraction of the energy usage, densities increasing, and hopefully a reversal in the recent trend towards less durability, SSDs will probably also overtake platter drives in price per terabyte within 5 years.

Re:SSDs will outpace platter drives

[jedidiah]

> By 2020, SSDs will have greater capacities than 20TB.

I will believe it when I see it.

Even with spinning rust you see a lot of vaporware and delay.

Gleeful declarations of the death of the other option are entirely premature.

> and hopefully a reversal in the recent trend towards less durability

That's the real kicker. I can bitch and moan about my Barracudas all I want but at least they give me ample warning before giving up the ghost. SSD has yet to prove itself in this regard.

Re:SSDs will outpace platter drives

[Mad Merlin]

With a fraction of the energy usage, densities increasing, and hopefully a reversal in the recent trend towards less durability, SSDs will probably also overtake platter drives in price per terabyte within 5 years.

It's not so much a trend as it is an unavoidable problem with increasing densities of SSDs. As you shrink the process size to fit more stuff in the same space, the durability goes down. As you change from SLC to MLC to TLC to fit more stuff in the same space, durability goes down. Substantial technological advances will be required to produce a 20T SSD with both acceptable durability and cost.

In comparison, shortscreen monitors (often mislabeled as widescreen) are a trend which has no logical or technical underpinning.

Re:SSDs will outpace platter drives

[Kjella]

There's still a >10:1 cost difference between 4TB HDDs and 1TB SSDs and SSD prices are not dropping that fast. Current SSDs are already on the bleeding edge of processing technology with 16nm MLC so there's fairly limited density increases and big durability issues ahead. I guess the wildcard is 3D NAND, but much like going multicore for CPUs it's a substitute. However, they are taking over all normal end user uses in cell phones and tables and laptops, it's just the big bulk storage left.

Re:SSDs will outpace platter drives

[sexconker]

With a fraction of the energy usage, densities increasing, and hopefully a reversal in the recent trend towards less durability, SSDs will probably also overtake platter drives in price per terabyte within 5 years.

It's not so much a trend as it is an unavoidable problem with increasing densities of SSDs. As you shrink the process size to fit more stuff in the same space, the durability goes down. As you change from SLC to MLC to TLC to fit more stuff in the same space, durability goes down. Substantial technological advances will be required to produce a 20T SSD with both acceptable durability and cost.

In comparison, shortscreen monitors (often mislabeled as widescreen) are a trend which has no logical or technical underpinning.

You could easily get 20 TB of flash today if you used a full size drive enclosure instead of the 2.5" wide, 7mm high shit they put out today.
I've got plenty of 5.25" drive bays. Give me a 5.25" SSD with a HDL, SATA Express/NME, or PCI-e 3.0 (x4 at least) connection. Then give me a chipset that lets me RAID across PCI-e storage devices, making sure I can boot off the resulting array and pass TRIM to the drives behind it.

Or how about someone starts making RAM drives (for a decent price!) again?

Re:SSDs will outpace platter drives

[BenJeremy]

I have some relatively ancient OCZ Vertex drives that are still running, 24/7 as OS drives for two of my servers (media and an ESXi box). Meanwhile, I have a large stack of platter drives that gave up the ghost with no warning whatsoever.

Reliability is as much a quality issue with SSDs as they are with platter drives, but there is less tolerance, more failure points with a platter drive, due to mechanical action.

Re:SSDs will outpace platter drives

[BenJeremy]

SSDs have far more room for innovation than platter drive technology. There are lots of promising advances making their way to production. Even better... you don't have to shrink the chips and make them more dense - you just have to make the existing fab cheaper. In 6 years, those chips will cost a fraction of what they do today.

As for monitors, 1080p is the result of convergence between the television and the computer monitor. Like it or not, it has resulted in an unprecedented reduction in price. South Korean sellers have already pushed WQHD (2560x1440) monitors to the edge of mainstream, and we'll see 4K mainstreamed in another two years (once the HDMI update gets widely adopted). Video games have been the driving force for computer technologies, and 1080p was a sweet spot, with the latest generation consoles finally able to support it fully and computers managing with even low-end video cards these days.

Re:SSDs will outpace platter drives

[BenJeremy]

Why should we be excited by a $30 1TGB drive today? 5 years ago I was paying $50 for 2TB drives.

That's my point. Development has slowed on higher capacity platter drives for a number of reasons... our demand as consumers might have slowed, but the "cloud" continues to grow and demand storage, but cloud providers are willing to spend too much for enterprise-grade storage they need. Technology is certainly a stumbling block, but they've been talking about these advances for many years. The main reason for the delays and jacked up pricing was plain and simple greed. The Thai floods were a godsend to the industry, which saw prices plummet below $0.025/GB - and suddenly, they were able to charge 3 times the price for all the drives they had stockpiled (not unlike the Sumitomo explosion back in '94 that drove RAM prices to 5 times their previous prices overnight - when the epoxy resin Sumitomo made was in plentiful stock supplies and never was short)

So platter drive makers have sat back and reaped profits, instead of staying ahead of the SSD price/performance/capacity changes. By 2020, those lines will have crossed. We now see "Enterprise" class SSDs, so capacities WILL continue to rise, even if most consumers only need a 1TB or 2TB drive on their PC. Server farms running only SSDs will be a thing in the future. They may even end up being more durable than platter drives by 2020, and that will make it an easy choice for cloud providers, even if it comes with a slight price premium.

Re:SSDs will outpace platter drives

[nabsltd]

Or how about someone starts making RAM drives (for a decent price!) again?

It's easier to just buy more system RAM than to have specialized "RAM drive" hardware.

Then, if caching doesn't give you good enough performance, you can just use software to create a RAM disk.

Re:SSDs will outpace platter drives

[0123456]

And at current flash prices, your 20TB SSD would only cost $10,000.

To be fair, that's less than we were paying for 128MB RAM drives in the 90s.

Re:SSDs will outpace platter drives

[Anonymous Coward]

The problem isn't space. If space was an issue, they'd just make 3.5" or 5.25" SSDs. The ONLY problem is flash prices. You just couldn't make a 20GB SSD for under $10k.

You won't see cheap RAM drives anytime soon either. There's just MUCH more complexity to it (power hungry DDR3 controller, backup power and all). It also means complex board layouts and low densities, on top of being a niche product with a low volume of sales. It's not gonna happen in the next 10 years, sorry.

A decent 256GB SSD costs less than 16GB of cheapo DDR3 and the gap keeps growing. Flash is getting cheaper while RAM has been getting pricier.

Re:SSDs will outpace platter drives

[FlyHelicopters]

And if you want to put a 1TB RAM drive into a server?

Frankly, the speed of modern RAM makes SSDs look slow, there is something to be said for ponying up the money for that much RAM (or more)

Re:SSDs will outpace platter drives

[dotancohen]

In comparison, shortscreen monitors (often mislabeled as widescreen) are a trend which has no logical or technical underpinning.

You are holding it wrong. No, really.

I've got a longscreen monitor (widescreen turned 90 degrees). It fits 120+ characters on a line and shows 122 lines of code plus two additional lines for the VIM status line. The other widescreen monitor is running Firefox with the tabs on the side (Tree Style Tabs) and the KDE panel on the left side of the screen, not the bottom. This is the absolute best setup that I could imagine.

Just stop wasting the vertical space and make better use of the horizontal space. I was a 4:3 holdout for the longest time, now I love the widescreens.

Re:SSDs will outpace platter drives

[Twinbee]

I can't help but think Microsoft would get Windows to use the HDD for the swap file even if 1000 terabyte of RAM were available.

Re:SSDs will outpace platter drives

[nctritech]

Backups, backups, backups.

Re:SSDs will outpace platter drives

[nabsltd]

And if you want to put a 1TB RAM drive into a server?

Buy hardware that supports 2-4TB on the motherboard.

Regardless of the physical size of the server, the price for such a motherboard plus RAM is going to be less than the add-in card plus RAM. It might be a lot more for the add-in solution, as you might not get warranty support unless you buy RAM from the same company that sells the card. You might also have to decide how much RAM disk you want at time of purchase, while system RAM can be upgraded (to the limit of the motherboard) and you can easily decide the size of RAM disk via software.

Also, for Linux, add-in cards might not have drivers, while system RAM does.

Re:SSDs will outpace platter drives

[mister_playboy]

Indeed, Windows 7 does huge amounts of swapping even though I'm almost always using less than half my 16GB of RAM.

Linux runs fine with no swap, but Windows 7 seems to have assumptions that swap will be used at all times.

Re:SSDs will outpace platter drives

[sexconker]

The next generation of flash will not use a drive form factor at all. SAS (et al) are way too slow for flash anyway.

SATA Express. Cabled drives running off a controller with dedicated PCI-e lanes using a newfangled port. Mobos with support are rolling out now.
The standard scales with PCI-e generations, too, as well as the number of lanes you throw at it (I believe the spec only calls for up to 4 lanes, but you could fairly easily up it to 16 lanes).

Re:SSDs will outpace platter drives

[sexconker]

But I want to BOOT off of it and have everything on one drive/partition. This isn't 1992, I don't enjoy juggling partitions for no fucking reason.
I want battery-backed, NAND-backed, RAM speeds. Blitz it all into the drive's RAM at POST. Then the OS doesn't even have to know WTF is going on.
At shutdown or loss of power, dump it all back to NAND.
You could probably even extend the life of that NAND dramatically by ONLY writing to it when power has been cut. So even when you've shutdown/halted, you're pulling enough power (from a direct connection to the PSU) to refresh the RAM, and trickle charge the battery.

$$$$$$ I'm sure, but fuck it.

Re:SSDs will outpace platter drives

[sexconker]

Adding support for more RAM to a mobo would cost a lot more than adding support for more RAM to a dedicated RAM drive device.
The RAM drive manufacturer will likely seek higher margins due to lower volume, though, so who knows how it would end up.

A RAM drive can present itself as any storage device, you shouldn't need drivers.

The one benefit using system RAM gives you is you're connecting directly to the memory controller, so you don't have to cry about how PCI-e x16 is your bottleneck.

Re:SSDs will outpace platter drives

[CopaceticOpus]

I've been watching storage price trends for the past five years.

Cost per 10TB of storage:

• Jul 2009: Platter = $750, SSD = $28,125
• Jun 2012: Platter = $567, Flash = $8,200
• Nov 2013: Platter = $450, Flash = $5,417
• Today: Platter = $373, SSD = $3,750

SSD progress has been amazing. The price for SSD storage is now 10x that of platters, compared to 37.5x in 2009. The cost for a platter drive today per TB is 50% of what it was five years ago, but for an SSD it is only 13%! Does it look like SSDs are about to take over? Not so fast. The rate of advancement has been slowing every year, and meanwhile platter drives are adopting major new technology this year.

Using a simple historical price model, I don't expect we would see price equivalence by 2020. However, by the end of that year SSD's may only cost 2x that of platter drives, and platter manufacturers will be very nervous.

Cost predictions per 10TB of storage:

• Dec 2016: Platter = $275, SSD = $1560
• Dec 2020: Platter = $160, SSD = $328
• Oct 2023: Equivalence. $109 for either technology.
• 2035: 10TB costs under $1 and is included in Happy Meals.

Re:SSDs will outpace platter drives

[Bengie]

Durability has already been solved tech wise via several different methods, now they just need to figure out how to incorporate it into their mass production.

Re:SSDs will outpace platter drives

[Bengie]

HP is working on a "Unified" memory where both the memory and storage is one massive pool of mram. The first gen servers will use partitioning to break up the pool between storage and memory, but later the OS will dynamically map pages on demand. They're expecting to start selling in the next several years, the tech is ready.

Re:SSDs will outpace platter drives

[toddestan]

Windows is constantly paging ram contents out the disk. That way, if something suddenly has a need for a large amount of memory, Windows can let it have it instantly because the contents have already been paged out. Linux doesn't do this so if something needs a lot of ram you have to wait while the kernel swaps something out to disk first. This is usually a good thing if you are trying to run Windows on a system without a lot of ram, but the thing is Windows keeps on doing it even if there is gobs of ram available. And if you think Windows 7 is bad about it, Vista was a lot more aggressive when it came to this.

Crossfire Hurricane

[Jeremiah Cornelius]

Who needs Terabyte Flash
When you spin rust in gas?

Re:Crossfire Hurricane

[Anonymous Coward]

As opposed to static in rocks?

Re:Crossfire Hurricane

[sexconker]

BURMA SHAVE

Just wondering...

[s0litaire]

Ain't Helium leakage only an issue under positive atmospheric pressure?
If the keep the Helium slight under normal atmospheric pressure it should stay inside the drive.
As long as the seal around it is good enough to keep other larger molecules out, the Helium will sit happily inside the drive..

Or am i missing something? never paid attention to those bits of science classes when i was younger...

Re:Just wondering...

[Em Adespoton]

You're missing the part where He is an inert monatomic gas -- since any pressure is created by how much helium you've got in that otherwise-vacuum, you'll find that the helium atoms, which can pretty much go anywhere they'll fit, will eventually find their way through the mass of other atoms bit by bit. They're small enough that they'll drift right through any "air pressure" as they can comfortably fit between most other gas molecules no matter how tightly they're packed together.

So it really all comes down to the seal: if they can get the seal to leave a gap no greater than two protons thick (He comes in stable isotopes of 1 or 2 neutrons), then no helium can escape. Good luck getting a seal that good though.

The truth is that statistically, most of the gas should stay inside the drive, but there will be constant loss, and eventually it will be enough that the gas cushion won't be enough to protect the platters and heads from collision.

Re:Just wondering...

[jeffb (2.718)]

So it really all comes down to the seal: if they can get the seal to leave a gap no greater than two protons thick (He comes in stable isotopes of 1 or 2 neutrons), then no helium can escape. Good luck getting a seal that good though.

Well, you just need to squeeze your neutronium together really hard along the joints.

Seriously, "a gap no greater than two protons thick"? Have you completely forgotten about electrons? You know, those things that hold all Earthly matter together (and apart)?

Re:Just wondering...

[Em Adespoton]

Electrons have a charge, but don't really take up physical space (and have no mass, therefore no gravitational pull). I was going to mention something about ionization, but the point of inert monatomic gasses is that they don't give up their electron shell, and don't steal from other molecule's electron shells. As such, the electron/proton binding is neutral, and so doesn't affect the surrounding materials.

Oh yes, and that whole thing your school teachers taught about electrons orbiting a cluster of protons and neutrons is a lie; it's just a convenient way of visualizing what's happening.

Re:Just wondering...

[ChrisMaple]

Atoms are about 10^-10 radius, protons about 10^-15. If proton diameter were the limiting factor in helium flowing through (for instance) iron, iron could no more contain helium than a fish net contains water. Tain't so; your explanation doesn't hold water either.

Re:Just wondering...

[jeffb (2.718)]

Oh yes, and that whole thing your school teachers taught about electrons orbiting a cluster of protons and neutrons is a lie; it's just a convenient way of visualizing what's happening.

Nice condescending swipe. Now, would you care to explain why you said you need "a gap no greater than two protons thick" to block the escape of helium atoms, each consisting of a nucleus with its attendant populated orbitals, several orders of magnitude larger than the bare nucleus that you seemed to be describing?

For that matter, how exactly would you define "a gap no greater than two protons thick" in an object made from molecular matter -- that is, matter bound together by those clouds of electrons that you alluded to? You know, the things that "don't really take up physical space" (except that they really do) and "have no mass" (except 9.10938291 × 10e-31 kilograms), and don't really "orbit" (but certainly do interact to form what's "conveniently" conceptualized as a van der Waals surface)?

Helium eh?

[Chas]

When approached for a comment, Dejah Thoris would only say "How have I come to be on Jasoom? And where is John Carter?".

Neato

[rsilvergun]

except that we've done away with the national Helium reserves and fraking + excessive natural gas mining is venting it all into space. There's a couple scientists that raised the alarm, since our entire tech is based on the stuff (and no, we can't just make more, and mining on the moon is _hard_). On the plus side I'll be dead by the time it's a problem.

why not build it stronger

[Ralph Spoilsport]

and put a vacuum in it?

Re:why not build it stronger

[xlsior]

and put a vacuum in it?

Wouldn't work: The read/write heads are actually floating a microscopic distance above the platter on cushion of air/helium/whatever. Without the gas, the distance between the heads and the platter would vary wildly, and it would almost immediately and literally come to a grinding halt, scratching up the disk surface in the process.

Re:why not build it stronger

[Hamsterdan]

Like Deskstars?

Re:Does HGST..

[slaker]

They no longer provide MBTF metrics for consumer (Deskstar) class drives but the uncorrectable error rate on Ultrastars is an order of magnitude higher than the others, which is a common distinction between consumer and enterprise-grade drives.

Re:Sounds stupid.

[rahvin112]

Do you have any idea how expensive RAM is? It's a couple factors larger than even NAND storage in price. If you want $20,000 hard drives your idea sounds great but the fact is RAM Is far too expensive. NAND and it's possible future replacements for non-volatile storage are the future. Spinning rust long term isn't going to be viable.

More importantly: Get Perpendicular follow-up!

[Brama]

Certainly we need an update to this classic: https://www.youtube.com/watch?v=xb_PyKuI7II

Do not want

[Trogre]

So with the slow yet inevitable leakage of helium, what will be the estimated lifetime of these drives?

Planned obsolescence, anyone?

Re:Do not want

[Neil Boekend]

Helium leaks slowly through a fully welded vessel. Really slowly.
If they use fully welded vessels the leak rate is around 10^-12 mbar*l/s. That means half the helium will be gone in approx 5 million years.

Re:typo??

[swamp boy]

Yes, it's a typo. It should say shingled (like shingles on a roof there is overlap).

Deja vu all over again

[pjwhite]

This has a familiar ring about it. We discussed this same story two years ago.

http://hardware.slashdot.org/s...

Not the first helium fillled drive

[dsgrntlxmply]

In the late 1960s, DDC of San Diego made head-per-track disk drives that operated with a helium atmosphere. These units had a cylinder of helium fastened to the baseplate (the units were 19" rack mount), and the documentation included procedures for replacing the cylinder and for purging from a full-sized cylinder if it was ever necessary to open the unit for repairs.

I had driven down to San Diego circa 1978 to buy a cylinder of refill helium from DDC for one of these in a hand-me-down system, but never got around to replacing the cylinder on the drive. The cylinder sat in my garage for years. Thirty years later I was a returned adult physics student. My professor was using a similar helium cylinder to purge a cryostat for a superconducting magnet. He ran out of helium, and the department had no other helium. I told him "wait 20 minutes, I'll be back." I retrieved the cylinder from my garage, and the professor was both delighted and baffled. When connected to the regulator, the cylinder proved to have maintained a remarkable fraction of its original pressure, and the professor was able to complete his procedure. Sadly, another part of the magnet failed and suffered a gas pressure explosion as it was being cooled.

In a remarkable coincidence, I noted that the department's helium cylinder and mine were identical, all the way down to a part number stenciled on them.

Lighter Laptop

[locopuyo]

How many of these do I need to get my laptop down to 0 lbs?

Aluminum leaks Helium.

[Hamsterdan]

Instead of an all-out pissing contest, can we have *reliable* 1 TB drives? Every drive bigger than 500GB has developped problems here a couple months past their warranty. Besides, Helium leaks unless properly sealed.

10TB of RAM?

[jeffb (2.718)]

You seem to have a really... optimistic view of the size, cost, and power budget for RAM.

Re:10TB of RAM?

[jd]

Not really. RAM is only expensive because of the transistor size used. Fab plants are expensive. Packaging is expensive. Shipping is expensive. Silicon is expensive. If you add all that up, you end up with expensive products.

Because fab plants are running very large transistor sizes, you get low yields and high overheads.

Let's see what happens when you cut the transistor size by three orders of magnitude...

For the same size of packaging, you get three orders of magnitude more RAM. So, per megabyte, packaging drops in cost also by three orders of magnitude.

Now, that means your average block of RAM is now around 8 Tb, which is not a perfect fit but it's good enough. The same amount of silicon is used, so there's no extra cost there. The shipping cost doesn't change. As mentioned, the packaging doesn't change. So all your major costs don't change at all.

Yield? The yield for microprocessors is just fine and they're on about the scale discussed here. In fact, you get better. A processor has to work completely. A memory chip also has to work completely, but it's much smaller. If the three round it fail testing, it doesn't affect that one. So you end up with around a quarter of the rejection rate per unit area of silicon to a full microprocessor.

So you've got great yield, same overheads, but... yes... you can use the fab plant to produce ASICs and microprocessors when demand for memory is low, so you've not got idle plant. Ever.

The cost of this memory is therefore exactly the same as the cost of a stick of conventional RAM of 1/1000th the capacity.

Size - Exactly the same as the stick of RAM.

Power budget - of no consequence. When the machine is running, you're drawing from mains power. When the machine is not running, you are refreshing the dirty bits of memory only, nothing else. And 99.9% of the time, there won't be any because sensible OS' like Linux sync before a shutdown. The 0.1% of the time, the time when your server has been hit by a power cut, the hard drive is spun down to save UPS and the main box is in the lowest possible energy mode, that's when this sort of system matters. Even on low energy mode, buffers will need flushing, housekeeping will need to be done, transactions will need to be completed. This system would give you all that.

And the time when the machine is fully powered, fully up? Your hard drive spends most of its time still spun down. Not for power, although it'll chew through a fair bit - mechanical devices always do and the high-speed drives being proposed will chew through far, far more. They'll be spun down because a running hard drive suffers rapid deterioration. Can you believe hard drives only last 5 years??! Keep the damn thing switched off until last minute, then do continuous write. Minimizes read head movement (there's practically none), minimizes bearing wear-and-tear, eliminates read head misalignment (a lot of times, you can write the entire disk in one go, so what the hell do you care if the tracks are not perfectly in line with the ones they're replacing?) and (by minimizing read head time over the drive) minimizes the risk of a head crash.

I reckon this strategy should double the expected lifetime of drives, so take the cost of one 10 Tb drive and calculate how much power you'd need to consume extra for the memory in order for the memory's power budget to exceed the value of what you're doing.

Oh, and another thing. Because I'm talking memory sticks, you only need to buy one, subsequent drives of the same or lower capacity would not need to have memory there. You could simply migrate it. RAM seems to hold up ok on old computers, so you can probably say that the stick is good for the original drive and the replacement. That halves the cost of the memory per drive.

So, no, I don't see anything unduly optimistic. I think your view of what the companies could be doing is unduly pessimistic and more in line with what the chip companies tell you that you should think than what the chip companies can actually do.

Re:10TB of RAM?

[jeffb (2.718)]

Let's see what happens when you cut the transistor size by three orders of magnitude...

Oh, is that all you have to do?

That was an awfully big wall of text to write for just one nibble.

Re:10TB of RAM?

[jd]

Memory transistors are about a thousand times larger than CPU transistors. Do try to keep up.

Re:10TB of RAM?

[jeffb (2.718)]

I can't possibly hope to "keep up", because you'll always be able to make up random nonsensical claims faster than I can debunk them. Enjoy your perpetual leadership.

Re:Skpetical

[Hamsterdan]

Thanks, I'm thinking the same thing. 5 years down the road those drives won't work. I'd rather have reliable drives than a dick contest...

Re:Skpetical

[Hamsterdan]

They will work for 6 months then fail in a spectacular way (and lose all the data in the process)

Myanmar much?

[tepples]

If these signs / were here today / the final one / would likely say / Myanmar Shave

Helium is non renewable

[gelfling]

And fairly limited. So.....good work.

Re:Helium is non renewable

[50000BTU_barbecue]

I'll bet there's going to be a small market for Fusors to create your helium on-site!

Re:Helium is non renewable

[dsgrntlxmply]

In second semester undergrad physics, one of our homework problems was to examine the tail of the Maxwell-Boltzmann distribution for helium at ordinary temperatures, relative to Earth escape velocity.

Nothing can replace helium, and once released to the atmosphere, it will 1) become too dilute to economically recover from the atmosphere, and 2) eventually escape into space.

Nothing especially (apart perhaps from magnetic and optical manipulations in very limited circumstances) can replace 3He as an ultimate refrigerant. Some 3He comes from natural He reservoirs, but most comes from 3H (tritium) production and decay. The decline of tritium production for nuclear weapons, has brought a corresponding decline in availability of 3He. One of my professors was in a dispute with his former institution over who owned a small cylinder of 3He from his former lab. It was at the time valued at around $16000, and was languishing uselessly in some provost's office.

7200 RPM spindle speed

[Eunuchswear]

Who uses 7200 RPM in the data centre?

We don't even have any 10K drives any more, they're all 15K, being phased out for SSD's.

Re:Hard drives are ludicrously amazing

[Eunuchswear]

Once upon a time there were 20MB hard drives. For real.

20MB! We used to dream of 20MB!

Not a joke - the first machine I used, started out with 8MB drives. Each the size of a washing machine.

It was eventually upgraded to 60MB drives.

(The machine was an ICL 1903T).

I think I can, I think I can...

[Thud457]

Once I complete my dirigible yacht, surely I can win the heart of the Princess of Helium.
Not that lousy Disney one either, but the proper nekkid one.

putting the cart before the horse

[slashmydots]

In case you haven't stopped by Newegg lately, look at the reviews for all 1TB drives. It's like 95% positive. Look at the reviews for any model 2TB drive from any manufacturer. Across the board they're HORRIBLE! It makes it seem like the DoA numbers are between 50% and 75%. Some people said they ordered 2, got them both DoA, then got 1 DoA out of the two replacements. Tons of other people claim that they failed within 3 months. Now look at the 3Tb and 4TB drives. Horrible again!

So apparently after making 1TB drives, Seagate and WD forgot how to make hard drives. I wouldn't buy one of these 8TB or 10TB drives for any reason because I expect the exact same thing to happen to them.

I shudder to think of a 10TB drive resilvering.

[bareman]

We just had a 3TB drive fail and it took 172hours to resilver the drive. [ZFS Raid 6]. That was long enough that we're switching to Raid 10. I wonder if a 10TB drive could rebuild before too many drives fail and take out the array.

Helium?

[sudon't]

You guys are discussing the helium reserves? I want to know how long it'll be before I can get one on Amazon for $99.

Ne or Ar instead?

[servant]

Yes, I know Neon and Argon are 'bigger' atoms, and weigh more, but they too are inert, and if lowering the pressure in the cavity, they could use either to help, and it would help preserve Helium reserves. Currently folks are clamoring to not use Helium for 'recreational balloon' purposes, because the He supply is evidently running low. The biggest use? NMR machines, the use it to super cool their magnets and seem to loose the helium at a pretty astonishing rate. Fixing that would save the world a LOT of He.

Re:Sounds stupid.

[jd]

I've a very good idea that RAM prices are artificially inflated, that the fab plants are poorly managed, that the overheads are unnecessarily high because of laziness and the mentality in the regions producing RAM.

I'm absolutely certain that 15nm-scale RAM on sticks the same size as sticks used today would cost not one penny more but would have a capacity greater than I've outlined.

It could be done tomorrow. The tools all exist since the scale is already used. The silicon wafers are good enough, if they can manage chips 4x and 9x the size of a current memory chip with next to zero discards, then creating the far smaller dies (so you can discard more chips and still get the same absolute yield) is not an issue. It would reduce idle time for fabs, as fabs are currently run semi-idled to avoid the feast/famine cycle of prior years but 15nm would let them produce other chips in high demand, soaking up all the extra capacity.

What you end up with is less waste, therefore lower overheads, therefore higher profit. The chip companies like profit. They're not going to pass on discounts, you getting a thousand times the RAM for the same price is discount enough!