Seagate Releases 6TB Hard Drive Sans Helium

Lucas123 (935744) writes "Seagate has released what it said is the industry's fastest hard drive with up to a 6TB capacity, matching one released by WD last year. WD's 6TB Ultrastar He6 was hermetically sealed with helium inside, something the company said was critical to reducing friction for additional platters, while also increasing power savings and reliability. Seagate, however, said it doesn't yet need to rely on Helium to achieve the 50% increase in capacity over its last 4TB drive. The company used the same perpendicular magnetic recording technology that it has on previous models, but it was able to increase areal density from 831 bits per square inch to 1,000. The new drive also comes in 2TB, 4TB and 5TB capacities and with either 12Gbps SAS or 6Gbps SATA connectivity. The six-platter, enterprise-class drive is rated to sustain about 550TB of writes per year — 10X that of a typical desktop drive."

Comments

~1000 *Bits* per square inch?

[PaulBu]

I thought that in 21st century we are talking about Gbits/inch^2, not just bits...

Paul B.

Re:~1000 *Bits* per square inch?

[cheater512]

Its just sad that in the 21st century we are still using Gbits/inch^2 and not Gbits/cm^2.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Pishposh. The hillbillies of the world with their crazy units again. How many nibbles does it store per library of congress floor tile?

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Its just sad that in the 21st century we are still using Gbits/inch^2 and not Gbits/cm^2.

Why is that sad? An inch is 2.54 cm, so a square inch is 2.54*2.54=6.45 cm^2. If we switched from inches to cm, instead of 6 TB, this disk would not even hold 1 TB.

Re:~1000 *Bits* per square inch?

[nmb3000]

I thought that in 21st century we are talking about Gbits/inch^2, not just bits...

Paul B.

That caught my eye as well. Assuming 1000 bits per square inch, we're talking about:

6 * 1000 * 1000 * 1000 * 1000 * 8 / 1000 = 48,000,000,000 square inches to store 6TB at 1000 bits per in^2.

1 Mbit per square inch makes a lot more sense, putting it at 48 square inches, or about 8 square inches per platter.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Yes, very sad we continue to use outdated and imprecise forms of measurement. We should stop mucking about and start counting moles, atoms, and elementary particles.

Re:~1000 *Bits* per square inch?

[Xicor]

... that statement is just plain dumb. bytes of data are the same 8bits regardless of storage density. if they switched it to cm^2, they would just change the number from 1000/inch^2 to 1000/2.54/cm^2

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Dammit Sir,
  Don't you know in the 21st century we use Libraries of Congress per Football Field.
or for small denominations, sometimes the Relevant Slashdot Submissions per Postage Stamp.

Re:~1000 *Bits* per square inch?

[SpankiMonki]

Better check your math on that. It's actually 1Tb/inch^2.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Whoosh!

And you got the unit conversion wrong! Gold!

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Imperial or metric bits?

Re:~1000 *Bits* per square inch?

[nmb3000]

1 Mbit per square inch makes a lot more sense

Oh, derp. Make that 1000 Gbit per square inch. Worst part of no edit on Slashdot is all the simple math mistakes irrevocably left for posterity :)

Re:~1000 *Bits* per square inch?

[radiumsoup]

^ winner.

Re:~1000 *Bits* per square inch?

[BoRegardless]

Slashdot is hiring fired AP writers now?

Re:~1000 *Bits* per square inch?

[Smerta]

Dude, sorry but you just got tripped up by Muphry's Law.

Happens to all of us, sooner or later...

Re:~1000 *Bits* per square inch?

[cheater512]

And that is the perfect example of why we aren't using Gbits/cm^2. You went for it hook line and sinker and said something *really* stupid. :P

Re:~1000 *Bits* per square inch?

[Xicor]

yea... i noticed after i posted, but sadly there is no edit..

Re:~1000 *Bits* per square inch?

[durrr]

Who cares, at least my music will no longer have a funny pitch on playback.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Why do you folks who put your name as your ID feel the need to use the same damn thing as your sig?

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

After that comment, I think your "Funny" mods above are undeserved.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

whooooosh!

Re:~1000 *Bits* per square inch?

[Ghaoth]

Yes, but the disk is rotating at 198,374 furlongs/fortnight.
It's all getting a bit bitty methinks.

Re:~1000 *Bits* per square inch?

[rossdee]

You don't get into the imperial or metric part until you're talking Kilo bits

Metric Kilo = 1000
Imperial Kilo = 1024

Re:~1000 *Bits* per square inch?

[davester666]

Imperial bits automatically reject the force.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

when the richest country (the one that buys everything) says its one set of units, it doesn't matter what other units exist, the units the richest country wants are the ones used.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

what about cubits? why isn't anyone using them? (if you don't get the joke there, this place is worse than I thought.)

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

Yes, very sad we continue to use outdated and imprecise forms of measurement. We should stop mucking about and start counting moles, atoms, and elementary particles.

Its very sad that people worry about truly inconsequential things like Gbits/cm^2 when we could call it 1*10^13b/(m^2s)

Re:~1000 *Bits* per square inch?

[peragrin]

how can you have 8 square inches on a round platter?

if the inches are square then they can't be round at least at that pixel density.

Re:~1000 *Bits* per square inch?

[Anonymous Coward]

1 Mbit per square inch makes a lot more sense

Oh, derp. Make that 1000 Gbit per square inch. Worst part of no edit on Slashdot is all the simple math mistakes irrevocably left for posterity :)

That is 1000 Mbit. Perhaps you should just stop posting.

Re:~1000 *Bits* per square inch?

[TeknoHog]

Is that an African or a European bit?

Let me just

[CSMoran]

write the partition table.

Re:Let me just

[Anonymous Coward]

I don't even bother with partition tables, just raid6 the entire block device.

Re:Let me just

[Anonymous Coward]

You think you could afford enough of these to facilitate raid 6, eh? Lol.

Re:Let me just

[MightyYar]

RAID 6 is only one more than RAID 5.

Re:Let me just

[Anonymous Coward]

You think you could afford enough of these to facilitate raid 5, eh? Lol.

Re:Let me just

[Anonymous Coward]

given probability of failure during recovery with raid5, you can't afford to not raid6 such large drives

Re:Let me just

[schitso]

Yeah, I pity anyone who runs RAID5 on such large disks. Enjoy your UREs.

Re:Let me just

[sexconker]

RAID 1+5 is the answer. 6 6 TB disks disks gets you 12 TB (which is really about 10.9 TB).

A1 A2
B1 B2
P1 P2

A1 and A2 are mirrors of each other (RAID 1).
A, B, and P form a RAID 5 array.

If A1 goes down, it gets rebuilt from A2.
If A2 goes down while rebuilding A1, you can recover A1 from B1 and P1.
If B1 or P1 then goes down, you can recover both it and A2 (and thus A1) from B2 and P2.

You can survive any 3 drive failures. You can potentially survive a 4th drive failure, too:

XX XX
XX B2
P1 P2

If P1 goes down, you can still rebuild everything automatically.
If P2 goes down, you can still rebuild everything, but your controller almost certainly won't automatically do so. You'd want to carefully read the documentation / contact your manufacturer to make sure you've got the right procedure for using P1 and B2 to regenerate A (either as A1 or A2).
If B2 goes down, you're fucked.

Even in the cases where you're fucked you may be able to get shit back together. Unrecoverable read errors are the most likely issue to come up when rebuilding an array when a disk goes down, but UREs are unlikely to line up across both the RAID 1 and RAID 5 arrays. If your shit looks like this (where X is unreadable data and O is readable data):

A1 [OOOOOOOOXXOOOOOO] A2 [XXXXXXXXXXXXXXXX]
B1 [OOOXXOOOOOOOOOOO] B2 [OOOOOOOOXXOOOOOX]
P1 [OOOOOXOOOOOOOOXO] P2 [XOOOOOOOOOXXOOOO]

Then you can still recover shit if you really want, though it's unlikely for your controller to be able to do so automatically while online (or automatically at all).

Compared to a single disk, RAID 1+5 is sextuple the cost for double the storage, but you get a ton of redundancy.
Compared to RAID 1+0, you get the same storage for 50% more money, but you can withstand double the amount of disk failures.
RAID 1+0 can survive 1 or 2 disk failures (2/3 chance for the second failure). RAID 1+5 can survive 3 or 4 (1/3 chance for the 4th failure).

Re:Let me just

[hobarrera]

Of course, instead of going bankrupt after spending all my money in 6TB hard drives, I could just enjoy myself for now, and got bankrupt later on when one of by four 3TB disks fail.

Re:Let me just

[sexconker]

If that's bankrupting you you likely don't have data worth replicating that much.
Besides, the point is to spend $XXXX on RAID instead of $XXXXX on data recovery or $XXXXXX on lost business / court settlements.

Big Drive

[Anonymous Coward]

6 TB drive... 1000 bits/square inch... So... 6 billion square inches of real estate (14.67 mile sides of a square)?

Re:Big Drive

[war4peace]

Also, "increase in capacity over it's last 4TB drive".

1000 bits per square inch!

[Anonymous Coward]

It's like magic!

oops

[Anonymous Coward]

Whoa, the summary is orders magnitude off on the density. (or the drive is way bigger than an aircraft carrier.)

Re:oops

[SpankiMonki]

Whoa, the summary is orders magnitude off on the density. (or the drive is way bigger than an aircraft carrier.)

So is the info in the forwarded Computerworld article. As this is a first gen HAMR drive, the density is actually 1 terabit per square inch.

Let the 4K porn revolution commence!

[Anonymous Coward]

North's law is the observation that, over the history of computing hardware, the number of bytes on magnetic storage doubles approximately every two years. The law is named after legendary porn star Peter North.

Why not?

[wisnoskij]

And why would you not use helium? They already seal the hard drives and it is just as easy and cheap to leave helium in the drive as some form of super clean air.

Re:Why not?

[Anonymous Coward]

No. drives are *not* sealed. Making a sealed drive that won't implode if you, say, take it on an aircraft in your laptop, or to ship it to the client (for example) is non trivial.

Re:Why not?

[Anonymous Coward]

Not as cheap. Industrial helium costs for grade A helium are incredibly high, +1 for them finding a way to do it without driving up a cost on something that doesn't have to happen 'yet'.

Re:Why not?

[Anonymous Coward]

actually, the drives aren't usually sealed. there's a porous frit plug that allows the pressure inside and outside to be equalized, and which has pores small enough not to let water vapor in.

Designing a hermetic container that lasts for years is non-trivial, particularly one to hold Helium, which has very, very small atoms that can go through the interstices in the metal. By comparison, pulling a vacuum would be easier, but disk drives need something to fly the heads on. He is nice because it is also very low viscosity

Re:Why not?

[pushing-robot]

Because they're trying to maintain their quality.

(Seriously, though, they actually don't seal their hard drives. Hard drives typically have a small hole in the casing with a extremely fine dust- and moisture-proof filter attached. It allows the drive to equalize its pressure with the environment and reduces the mechanical strain on the housing and seals.)

Re:Why not?

[Kaenneth]

Fun Fact, retail helium for recreational used is often salvaged from used 'pure' helium from MRI machines and such.

So children's party balloons are filled with medical waste, yay!

Re:Why not?

[niftymitch]

And why would you not use helium? They already seal the hard drives and it is just as easy and cheap to leave helium in the drive as some form of super clean air.

Ask the question differently..... why you would use helium.
A partial answer is thermal... revisit the old TCM (thermal conductive modules)
used in system components like IBM 3081. H has good thermal properties.
Another is the H or He have vastly better dynamic fluid properties to let the head seek
and perhaps fly better.
Pressure and seal is an issue but an expansion bladder can establish a way for
pressure changes to have minimum impact on the case.
Helium is inert allowing a wider choice of internal materials.

Some of these properties are not dependent on a 100% H atmosphere.

And if you crack one open you can sound funny.

Re:Why not?

[war4peace]

He is nice because it is also very low viscosity

But she is nicer than him, that's for sure, mainly because her viscosity is spot on.

Re:Why not?

[niftymitch]

.......

Designing a hermetic container that lasts for years is non-trivial, particularly one to hold Helium, which has very, very small atoms that can go through the interstices in the metal. .....

This is important.....
I have seen thick wall steel pipe with blisters inside the metal.
The pipe contained high pressure hydrogen at high temperatures and
the hydrogen would react with the carbon in the steel and grow bubbles
a little like Swiss cheese.

The diffusion dimension of H2 and He is also interesting.

Re:Why not?

[Anonymous Coward]

Fun Fact, retail helium for recreational used is often salvaged from used 'pure' helium from MRI machines and such.

So children's party balloons are filled with medical waste, yay!

I wondered where Big Bang Theory got most of their scripts for Sheldon. Now I know ... Slashdot!

Re:Why not?

[Anonymous Coward]

+2 moron mods. Thanks for making a thousand unsuspecting wanna be computer guys a little stupider today.
Seriously, why don't you people that know nothing STFU.

Re:Why not?

[Anonymous Coward]

He also has the advantage of not turning your computer into a conflagration like H.
an expansion bladder?
a wider choice of internal materials???
holy shit.
I can tell you are one of those really smart guys that has never built anything in his life.
please
please
please
SHUT YOUR PIE HOLE

Re:Why not?

[Anonymous Coward]

Because they're trying to maintain their quality.

(Seriously, though, they actually don't seal their hard drives. Hard drives typically have a small hole in the casing with a extremely fine dust- and moisture-proof filter attached. It allows the drive to equalize its pressure with the environment and reduces the mechanical strain on the housing and seals.)

http://www.tweaktown.com/articles/6028/dispelling-backblaze-s-hdd-reliability-myth-the-real-story-covered/index.html

That comparison (and the community response) is like getting angry when you put a heavy equipment trailer behind a couple of compact cars and the transmission explodes on one faster than the other; in other words if you put desktop, or god forbid, drives that were destined to live there lives in consumer external drive enclosures, in some sort of commercial application, you're asking for a LOT of trouble, regardless of HD vendor.

Re:Why not?

[dargaud]

Designing a hermetic container that lasts for years is non-trivial

Huh ? My cheap plastic watch is hermetic to 4 atm, and for years. Plenty of things are. And for a HD you don't have to stand more than 1/3 atm of pressure differential, something trivial. Having used hard drives at high altitude and seen them die quickly, I always wondered why they don't simply seal the damn things with air at 1 atm inside.

Re:Why not?

[Anonymous Coward]

Surely that would be water resistant rather than helium-proof though. And water exists as rather large molecules compared to individual helium atoms.

Re:Why not?

[thegarbz]

it is just as easy and cheap to leave helium in the drive as some form of super clean air.

Have you ever wondered how Helium and Hydrogen filled balloons, even metal ones will deflate and fall despite no obvious signs of a leak? It is very difficult to properly seal either element into a permanent storage container. There's nothing cheap or easy about designing a case like that.

Re:Why not?

[JoeMerchant]

So, when you pull a vacuum, does it eventually fill up with helium that goes through the interstices in the metal from the ambient atmospheric mix of gasses?

Re:Why not?

[gmarsh]

Why not?

- Requires extra machinery/infrastructure to be installed at the HD plant
- Expense of buying helium
- Expense of other drive changes required to support helium (better sealing, etc)
- Expense of licensing existing patents on helium filled drives and their manufacture.

Or perhaps they're just busy validating new drive models and their assembly line, and they'll be shipping helium filled drives shortly...

Re:Why not?

[LoRdTAW]

Helium is so good at leaking it is used to test for leaks using a mass spectrometer leak tester.

Re:Why not?

[Linzer]

Yes, at an incrediby slow pace, and only until you reach the same partial pressure of helium inside and outside the container. Which will give you a ridiculously small amount of helium after a huge wait. That's why no one does that.

Re:Why not?

[TeknoHog]

No. drives are *not* sealed. Making a sealed drive that won't implode if you, say, take it on an aircraft in your laptop, or to ship it to the client (for example) is non trivial.

By "ship", do you mean a submarine? Because otherwise my head in plode (considering a roughly sea-level internal pressure vs. the mile-high club)

Helium leak?

[Tablizer]

Finally a way to detect if your drive is about to crash: you start to sound like a munchkin.

<highvoice>Oh oh, I better make a backup fast!</highvoice>

Re:Helium leak?

[Anonymous Coward]

What's wrong with using Acetylene? You'd be notified of a drive failure from the other side of the DC.

Re:Helium leak?

[Jamlad]

I'm wondering how they've manage to prevent He from leaking. It's an absolute bugger to contain. Just permiates through most plastics and any infinitesimally small cracks. We use it in particle accelerators at ultra-high vacuum (10^-9mbar) to detect leaks for that very reason.

Cloud Compatible

[Anonymous Coward]

Complete with humidity sensor.

Yeah, yeah. What's it cost?

[Anonymous Coward]

Been a long time since I've bought Seagate as I've had great reliability from WD but out of curiosity, what's the cost?

Re:Yeah, yeah. What's it cost?

[fnj]

Been a long time since I've bought Seagate as I've had great reliability from WD but out of curiosity, what's the cost?

We won't know shit until Newegg begins to stock them, maybe in about 5 years. Cost data from manufacturers, even when you can get it, is always useless because in the real world nobody pays anywhere near list price.

Calculations

[wisnoskij]

At 1000 bits per square inch, to get 6TB you need about a third the size of Manhattan.

According to Wolfram Alpha at least:
http://www.wolframalpha.com/in...

Re:Calculations

[steelfood]

The He-filled WD version is just a bit larger, and has a giant laser in the center that can blow up a city.

Re:Calculations

[Anonymous Coward]

I liked the 0.09-0.2x the average range of a male cougar better. Although I thought a male cougar was just called a pedophile...

Re:Calculations

[Anonymous Coward]

What I really want to know is how many stadiums would be required to store a library of congress.

Re:Calculations

[Trailer Trash]

At 1000 bits per square inch, to get 6TB you need about a third the size of Manhattan.

It takes a couple of hours to get up to 5800RPM, but when that bitch is spinning don't even try to tilt your computer.

Re:Calculations

[JoeMerchant]

Just think, last year's drives at 831 bits per square inch, that's only 29 bits per inch, squared... you can see those suckers with the naked eye.

Re:Calculations

[Hobadee]

If I'm doing my math right here, that comes out to ~1,900km/s at the outter edges of the platter. That's about 0.6% of the speed of light.

Wolfram Math ...Divide by pi and sqrt to get radius, multiply by 2pi to get circumference, multiply by RPM, divide by 60 to get it in seconds... Correct?

Re:Calculations

[Trailer Trash]

If I'm doing my math right here, that comes out to ~1,900km/s at the outter edges of the platter. That's about 0.6% of the speed of light.

Wolfram Math ...Divide by pi and sqrt to get radius, multiply by 2pi to get circumference, multiply by RPM, divide by 60 to get it in seconds... Correct?

So if it's a third the size of Manhattan at rest, how big is the outer edge at full speed?

"relying" on helium

[Anonymous Coward]

Seagate, however, said it doesn't yet need to rely on Helium to achieve the 50% increase in capacity over it's last 4TB drive

Seagate previously made 4 x 1 TB platters and 5 x 800 GB platters. Now this drive stores 1.25 TB per platter (according to El Reg). I bet WD/HGST can replicate that very easily... 7 x 1.25 = 8.75 TB. From what we know Seagate could use shingles (shingled magnetic recording) to boost capacity but with a penalty to write speed. There was also a suggestion they could cram 6 platters in a drive without helium. Both companies are working on HAMR to replace PMR in the coming years.

It's not that WD is relying on helium, it's that WD has better technology than Seagate. By including two more platters, WD can match Seagate's capacity with older PMR platters.

Re:"relying" on helium

[Anonymous Coward]

Who cares about who can cram more bitdicks in to more dickplatters. Which is more reliable?

Re:"relying" on helium

[Anonymous Coward]

Doesn't including more platters increase the chance of a hard drive mechanical failure?

Re:"relying" on helium

[Anonymous Coward]

fewer platter == fewer heads == fewer moving parts == more reliability.....assuming all heads/platters/etc. are created equal.

Re:"relying" on helium

[Anonymous Coward]

More platters = more spinning mass = more energy required = more heat. Heat is not a friend of hard drives. Plus more moving parts (actuator heads) = more points of failure.

As a general rule of thumb, 5 and 6 platter drives are worse then 4 platter drives in terms of reliability. That has been a good guideline going back 15-20 years.

Sorry, but Seagate is total shit

[Anonymous Coward]

And we all know it.

Re:Sorry, but Seagate is total shit

[Anonymous Coward]

I dont, in 27 years I have had 1 seagate fail, and it was still under warranty, meanwhile every single WD I have ever bought has shit its shoes and hit the dumpster

Re:Sorry, but Seagate is total shit

[gweihir]

WD has a know problem is sulfur-rich environments (e.g. former east-bloc) as they use silver-plating instead of gold on their PCBs. Other than that they are perfectly fine. Seagate is by far the most flaky vendor with often good disks and often lemons.

Re:Sorry, but Seagate is total shit

[Anonymous Coward]

I've had 4 of my 12 Seagate 3TB NAS drives die inside 6 months. I've been replacing them with WD RED drives with the idle timer set to 300s.

It's bad that WD started implementing an 8 second idle park timer on their RED drives. Luckily you can still modify this parameter with their DOS tool and set it to something reasonable.

Its, not it's

[Anonymous Coward]

it's last 4TB drive

NO, NO, BAD Unkown Lamer!

"Its last 4TB drive", as opposed to "It's..... Monty Python's Flying Circus!"

So with the Helium inside, is drive lighter?

[Glasswire]

In a big storage server, that could amount to few kilos, perhaps

Re:So with the Helium inside, is drive lighter?

[gargleblast]

In a big storage server, that could amount to few kilos, perhaps

Absolutely. That's astute! They're going for cloud storage.

Press blurb is junk

[m.dillon]

The press blurb is full of nonsense. Not one real performance statistic. Not one.

-Matt

Re:Press blurb is junk

[gweihir]

You ask for facts? These have been eliminated from marketing a long time ago, as they tend to get in the way. There are enough suckers that will buy this thing anyways, despite Seagates consistently bad reliability track record.

obsolete

[Billly Gates]

Does anyone buy platters of very slow piles of rust anymore? I use 2 ssds. Will never go back due speed and reliability. Seagates do not last long and latest ssds can do 70 tbs before failing

Re:obsolete

[iggymanz]

yes. as you will find out in a couple years, SSD are more reliable in the first 2-1/2 years, then they go to shit faster than hard drives after that.

Re:obsolete

[Harlequin80]

Hmmm Let me think about this

Intel Enterprise grade SSD 800gb - I can find them for about $1800
HGST Ultrastar - Enterprise grade with Helium - 6tb - $865.

I currently have a 21tb Nas machine in this office. I don't need speed I need capacity. The cost differential is MASSIVE. So yes people buy spinning platters all the time.

Re:obsolete

[WuphonsReach]

Does anyone buy platters of very slow piles of rust anymore?

For bulk storage (measured in terabytes and petabytes), platters of spinning rust are the only economical solution. So for a secondary storage SAN where capacity is more important then IOPS, you fill it out with spinning rust. A 4TB enterprise SATA 7200RPM drive is about $330 right now, or about $0.08/GB. The cheapest enterprise SSDs are about 10x-20x that price.

Rust is also better for drives (or tapes) used as backups. It has better shelf stability then a SSD. Most SSDs will start to lose data after a few months of being disconnected (maybe as few as 6 months). Barring mechanical issues, traditional magnetic media holds up well over the span of years (at least a decade in most case).

Now I just wish WD would come out with a 2TB 10k RPM SATA Velociraptor 3.5" drive...

Re:obsolete

[dbIII]

I'll link the above post next time you pretend to know what is going on in server rooms so that everyone can have a laugh.

Re:obsolete

[Anonymous Coward]

Some people need capacity at a low cost over speed.

Facebook for example archives billions of photos nobody will ever look at again, because among them are millions of photos that people will look at again, and it's impossible to figure out a-priori which ones will be.

At a recentish lecture they gave at Stanford, they said what they really want is write-once solid-state memory with zero standby power, and lower costs than magnetic disks.

If you think about it, the vast majority of non-cache storage is write-once, so the idea of large low-cost write-once capacity should be broadly appealing.

Re:obsolete

[jabuzz]

No, they are all spinning aluminium or glass that are coated with none iron based magnetic materials. I don't think iron based platters have existed for many many years.

That said your post is amusingly clueless as to real world bulk storage. Unless there is some major breakthrough there is at least another five years before SSD world wide production can match spinning hard disk capacity, and that assumes hard disks stand still. On the other hand expect 15k and 10k disks to disappear over that time frame. Shipments of the faster drivers are already dramatically down as people are replacing them with SSD.

Re:obsolete

[Anonymous Coward]

Please, do that. Ignorance needs nothing more than repeated public shaming.

Re:obsolete

[MachineShedFred]

Believe it or not, capacity still matters.

Re:obsolete

[WuphonsReach]

Yep,

* 7200 RPM drives are in the $0.08-$0.12 per GB range.
* 10k RPM are around $0.23-$0.30 per GB range
* 15k RPM are around $0.50-$0.57 per GB range
* Good enterprise SSDs are now down to $1.00-$1.55 per GB.

SSDs are definitely putting the squeeze on the 15k RPM drive market. Price difference is now only about 2x between 15k SAS and SSDs. So for any application where you would short-stroke the drive to get more IOPS but less capacity, SSDs edge out the 15k RPM drives.

Unless the prices on 15k RPM SAS drives drop by a factor of 2x or 3x, there's potential for enterprise SSDs to overtake them by next year.

H2 drive

[c4tp]

I'm just glad we've learned the lesson of the Ultrastar Hindenburg. Oh, the capacity!

(Too soon?)

Re:H2 drive

[Anonymous Coward]

For people who are too stupid to know that Helium isn't the explody gas.

Re:H2 drive

[c4tp]

Yep. Didn't intend to whoosh anyone here, I just thought going too deep into an explanation of the difference between He(lium) and H(ydrogen) would make the joke even less funny than it already was.

Re:H2 drive

[Walter White]

Why not Hydrogen? Must utility generators are hydrogen cooled due to hydrogen's low viscosity and high thermal conductivity.

I'm surprised no one said, yet...

[kauaidiver]

That's a lot of porn...

Get Perpendicular

[ArcadeMan]

The company used the same perpendicular magnetic recording technology that it has on previous models.

Just in case anyone missed it, here's the very technical video explanation by Hitachi about how perpendicular recording works.

Re:Get Perpendicular

[Anonymous Coward]

I'm still waiting for a HAMR time video from them.

Re:Get Perpendicular

[Anonymous Coward]

3technical5me

Rebuild time

[Anonymous Coward]

Would love to know what the rebuild time would look like for a disk of this size. Seems impractical.

Don't buy Seagate drives...

[gweihir]

Sometimes, just sometimes, they are on par with the competition reliability-wise. But many of their drives are lemons, far more than from other manufacturers and that has been a very long-term trend. Seagate just does not know or does not care to make drives reliable. Latest data:
http://blog.backblaze.com/2014...

This one is unlike to be any better in that regard.

Re:Don't buy Seagate drives...

[Anonymous Coward]

Sometimes, just sometimes, they are on par with the competition reliability-wise. But many of their drives are lemons, far more than from other manufacturers and that has been a very long-term trend. Seagate just does not know or does not care to make drives reliable. Latest data:
http://blog.backblaze.com/2014...

This one is unlike to be any better in that regard.

I swear, it seems if you can generate the right kind of FUD that seems to resonate with "storage nerds" biases, it will be keep being thrown around like that damn Wakefield vaccine paper. If I were building network storage (and yeah, I suppose I'm not Backblaze trying to do it on the super cheap), there's no way I would do something as stupid as go buy a bunch of external drives and rip the drives out and put them in server application. Try doing that with HGST, WD, Seagate or Toshiba branded products, and you're going to get bitten.

    Using a car analogy, it's like John Doe crying foul because his Camry fell apart quicker than his Sentra when he put a tow hitch on there and was using them routinely to pull heavy equipment. Those drives were NOT designed to be run full-bore in a server environment. If you were actually thinking about the problem (which the a lot of the johnny-come-lately "cloud/app" folks who want to reinvent everything don't seem to do), you should have come to the conclusion, "I need enterprise grade drives", NOT desktop hardware. I want to pull heavy equipment--I should buy a full-ton truck or a dump trucker, not a compact car. Using drives in this capacity is really an invalid metric, and doesn't apply to the average user, where the heads actually get parked once in a blue moon, and the PC isn't necessary always on (or at least hopefully not always writing). Also, if Seagate drives were actually a bad value proposition, why does Backblaze keep buying them?

Also:
http://www.tweaktown.com/articles/6028/dispelling-backblaze-s-hdd-reliability-myth-the-real-story-covered/index.html

Re:Don't buy Seagate drives...

[gweihir]

Nice attempt at diversion. What you say is complete BS though, and rather obviously so.

First, you always get bitten when doing storage. There is no way around that.

Second, "enterprise grade" disk hardware is a myth today. Backblaze recognized that and quite a few other people do it too. Sure, you can get things like lower vibration, but they do not matter to drive lifetime these days, only access time and that you can measure. Your argumentation relies on 20 year old tech and that is not being used in modern drives anymore.

And third, they bought Seagate a) before they had that data and b) why do you think they were bad value? They were far worse than the competition, but as there was a HDD shortage, they can still be better having than not having.

The "story" by tweaktown you link has all the hallmark of cleverly disguised corporate "spin", which aims to muddle the waters but does not actually disprove anything or even explain anything. It is either paid-for disinformation or the work of idiots that do not understand engineering basics. The data and explanation by Backblaze is solid though.

Re:Don't buy Seagate drives...

[Anonymous Coward]

Not to mention there has been plenty of well known "bad batches" (read shit drives that Seagate sold anyway) in the past few years, much more than any other manufacturer of drives. Who could forget the 1.5TB drives? Bad ARM controllers, firmware that would slam the heads, ever since Seagate bought Maxtor they've been shit, instead of bringing Maxtor up to Seagate quality it brought Seagate down to Maxtor levels of garbage.

Re:Don't buy Seagate drives...

[Anonymous Coward]

If you don't actually think there is a physical, and QA difference between drives destined for enterprise and nearline storage and what gets slapped in your mom's DVR, I highly recommend disassembling some of the drives (even the same model will tend to have extra parts for drives destined for nearline applications). Even the drives that get shipped today, there is still a significant difference (as all of our vendors have explained to us). I buy equally as many WDs, HGSTs and Seagates, and honestly save for things like "Deathstar" and "Moose", which happen from time to time, the failure probabilities are about the same. Honestly it's a step up from all the old IBM-centric junk we used to have to deal with, but I digress; but to say something as unrealistic (as Backblaze does) that things like temperature don't affect drive functionality, is again, just wrong. As an experiment, you should put any drive in 120-140 degree chamber and beat on it in parallel, and grab one out of another batch and run it at room temperature. You'll find the hot drive will fail in roughly half the time, again regardless of vendor.

Re:Don't buy Seagate drives...

[gweihir]

Keep kidding yourself. And yes, I have disassembled drives and I know what I am looking at with electronics. I also do understand accelerated aging for electronics and you do very obviously not. The relationship between temperature and failure rate is a simplistic model, that mostly held for old drives as ball-bearings in the spindles do follow it. Guess what, modern drives do not have them anymore and FDBs are a whole different story. There are a lot of other things to understand with heat-accelerated aging.

The Backblaze people do know what they are talking about. Of course, Seagate fanbois and paid trolls do not want to hear that.

Re:Don't buy Seagate drives...

[gweihir]

Indeed. Good Seagate drive models are roughly on par with the average competitor drive. The problem is that they have so many bad ones and do not seem to care that they are pushing bad drives with significantly reduced reliability to their customers. And while other hdd manufacturers have had the occasional bad model as well, with Seagate it is a pattern and a quite frequent occurrence.

Re:Don't buy Seagate drives...

[Anonymous Coward]

Horseshit.

I bought 8 "enterprise" drives to put in a storage device from one manufacturer, and they all failed within a year with SMART errors. I replaced them with 8 drives that were spec'd the same, but did not bear the "enterprise" name from a different manufacturer, and they're still purring away two years later in the same box, with the same (or more) load, in the same environment.

Yeah, small sample size, etc. But a 100% failure rate is still significant in a small sample size.

Re:Don't buy Seagate drives...

[jbo5112]

Google's report on drive failures said enterprise drives were no more reliable. I actually compared specs for a vendor's desktop and enterprise models of a disk. The only difference on paper was with the touted vibration handling circuitry, that I'm told is an extremely important feature you get with buying enterprise drives. It was optional in the enterprise drives and included in every desktop drive. Most shocking was when I tore apart a very expensive high-capacity Sun RAID array from 2005/2006. Sun built the array with Hitachi Deskstars instead of the enterprise Ultrastars. Aside from a Sun logo, it was the same drive model that a friend gave me out of his Dell desktop.

Oops in title - "sans" ?

[perpenso]

Whoa, the summary is orders magnitude off on the density. (or the drive is way bigger than an aircraft carrier.)

I think that you can't get past the title without an oops: "Seagate Releases 6TB Hard Drive Sans Helium"

Doesn't "sans" mean without?

Re:Oops in title - "sans" ?

[dylan_-]

Doesn't "sans" mean without?

Yes, that's because WD's 6TB Ultrastar He6 was hermetically sealed with helium inside, something the company said was critical to reducing friction for additional platters, while also increasing power savings and reliability. Seagate, however, said it doesn't yet need to rely on Helium to achieve the 50% increase in capacity over it's last 4TB drive.

At least, I'm sure I read that somewhere.

55TB / year?

[zdzichu]

Since when mechanical drives degrade during writes? Isn't that a SSD 's illness?
And 55TB / year for desktop drive sounds ridiculously low.

Re:55TB / year?

[Anonymous Coward]

It's 550 TB/year. Which is 18.3 MB/second. Most desktop disks I have used cap at around 50MB/second writes to a real file system (not just dd) - they need RAID for better performance.
Anyway, it's a pretty good rating for sustained writes if it's guaranteed to perform well and not crap out.

Re:55TB / year?

[AmiMoJo]

Writes do slightly degrade HDDs. The extra heat shortens their lifespan, but the main issue is that they need idle time to check and re-write existing data.

When idle HDDs will do a background surface scan where they check the integrity of each block. All blocks will have some amount of errors that can be recovered using error correction codes. Once a block reaches a certain threshold of errors it is considered bad and the data is moved somewhere else. The drive has a supply of reserve blocks for this purpose. SMART data can tell you what percentage of these spare blocks has been used, and if you see you are getting low or the rate of re-assignment starts to increase you know you need to replace the drive before it dies.

If you saturate the drive with writes constantly it doesn't have an opportunity to do this background surface scanning. If a block degrades too much before being copied by the surface scan you get a read error, and the moment you get a read error the drive is considered broken and returned under warranty.

Re:55TB / year?

[Anonymous Coward]

Moving parts wear out. All reads and writes in a hard drive require something to move.
Writes are usually a lot more intensive, phisically. You have to write, and then read to verify that your write worked properly. Sometimes more than once. You also have to power up the head that changes the magnetic domains on the platter. Reading (as far as I know) is completely passive.

Flash actually works the same way. Writes require lots of power, and often require many write/verification passes. Reading is more or less passive.

Unfortunately, Seagate's reliability is garbage

[cpm99352]

There was a time when Seagate was a gold standard, but nowadays certainly not. I would not trust a Seagate drive if someone gave it to me. A truly sad state of affairs. I just pity the consumers who don't know any better. Even worse that they're now tarnishing the Hitachi brand.

Re:Unfortunately, Seagate's reliability is garbage

[Anonymous Coward]

Hitachi is a subsidiary of WD

Re:Unfortunately, Seagate's reliability is garbage

[Anonymous Coward]

Not really.
Hint: What's the difference between a Hitachi HDS723030BLE640 and a Toshiba DT01ACA300?

Re:Unfortunately, Seagate's reliability is garbage

[Anonymous Coward]

There was a time when was a gold standard, but nowadays certainly not. I would not trust a drive if someone gave it to me. A truly sad state of affairs. I just pity the consumers who don't know any better.

Fixed that for you. All hard drives suck, and always have. Best you can do is plan around that fact.

Re:Unfortunately, Seagate's reliability is garbage

[Zeio]

True. Maxtor infected seagate, and their reliability is generally horrific.

http://blog.backblaze.com/2014/01/21/what-hard-drive-should-i-buy/

25% of seagate drives die within 3 years. Its utterly horrific.

Re:Unfortunately, Seagate's reliability is garbage

[Anonymous Coward]

Seagate has many types of drives. Without comparison, that statistic is meaningless. It could be some older less reliable low end drives had that statistic (25% failure within 3 years), but given a comparable drive from WD or another vendor, you would likely find similar results. At the very high end, reliability is again significantly different. Apples and oranges, you can't just throw a blanket statement over all their drives.

Re:Unfortunately, Seagate's reliability is garbage

[kmg90]

just bought a 1TB 7200 RPM 2.5" HGST (what Hitachi became in the WD acquisition) it's been almost 4 months of daily use in my PS3 and not a single problem.

fir57

[Anonymous Coward]

www.Anti-slash.0rg

Helium drives

[maroberts]

...work at a very high pitch!

Computerworld screwup

[Anonymous Coward]

The blurb quotes Computerworld correctly, but the Computerworld author was fuzzy headed or didn't think that the number was *WAY* low or doesn't understand that 'computer something technical stuff'. The aerial density of this drive is now at 1 Terabit per square inch, not 1000 bits per square inch (a mere difference of only 1 billion times).

My bad ...

[perpenso]

Ah, I missed WD. I thought the summary was describing the Seagate drive as being filled with He. Gotta stop that skimming through torrents of information ... oh wait, that's a different article.

Humidity sensor

[fgouget]

The new 7,200rpm, enterprise drive also has what Craig described as a "humidity sensor" that will allow it to continue functioning in humid environments.

Finally a drive that will work in the cloud!

Please proofread your posts.

[azav]

> over it's last 4TB drive.

over its* last 4TB drive.

        it's = it is

Learn this.

Re:Please proofread your posts.

[Anonymous Coward]

Please jump off a bridge.

We measure in..

[Anonymous Coward]

We measure in Qberts here.

Why Not?

[Anonymous Coward]

Its called the double-dildo effect.

- Anonymous Cowyard

mmmm pie!

[Anonymous Coward]

gots to have the pie.

Nice for 3.5...

[Torp]

Now where are my 2Tb+ 2.5" hard drives?