"Around 10 years back, we had 2TB 3.5" HDDs ... In the last decade, we have seen advancements in three different categories that have enabled a 10-fold increase in the capacity of HDDs while retaining the same physical footprint"
its not anywhere near the same cost, your 10 fold increase is propaganda and misinformation. nevermind that your cost numbers are not TCO. I wouldnt be surprised if its a sponsored article by a failing industry (not saying it is)
You don’t know what you’re talking about. Just look at capacities and their prices. Even a child can do that. When I see a 2TB drive that used to cost $600, and now see an 18TB drive that costs $600, I’m pretty impressed, and you should be too.
It’s getting more difficult to increase capacities as time goes on. The sophisticated technology used these days is much more complex. It’s pretty surprising that it can even be done at these prices. I remember the days when the first had for personal computers came out, many years ago. It was 3.125MB, and cost $3,125. In today’s dollars, that’s almost $6,000. There’s been steady progress ever since.
"When I see a 2TB drive that used to cost $600" What?? 2TB drives certainly never cost anywhere close to $600. The first one launched in january 2009 with an MSRP of $250, but the actual price in stores nearly immediately dipped well below $200. In july 2010, which is 10 years ago, they cost around $100, 1/6th of what you're quoting. During the summer of 2011, before the dreadful floods of fall 2011 in Thailand, they cost below $60, less than 1/10th of what you're quoting.
Oh haven't you learned by now to ignore azfacea and their ramblings? They're like those cooks with "the end is nigh" on every corner if anyone will ever listen.
@azfacea, don't worry, hard disks are fake news, they don't actually exist. They stopped existing the moment you started rambling how they'll be completely replace by SSD is a matter of months. Now we're just buying the for collections. ;)
Dude, I remember when I was buying a 1GB drive for $330 (I'm old, get over it.) the dealer went into the back of the room and brought out the biggest drive he had ever sold. It was 10TB, and it cost $10,000.
At the same time, at the newspaper I worked for, we had a pagination system (design the newspaper on computer) that ran on a Sun Workstation (the old pizza box models). It had a 1 gigabyte storage system that was larger than a washing machine and cost $100,000.
Actually, fake news is about sources that completely make up information, such as the moon landing was staged, or JFK isn't dead, etc. When it came up in 2016 it was about sites that would publish articles that Obama was going to institute martial law rather than leave, or that the Pope endorsed Trump, or that RuPaul accused Trump of something. All of those were fake news.
Inaccurate stories is something else. However, stating that we've seen a tenfold increase in capacity is either true or it's not. Not mentioning cost or other facts has no bearing on if there was a ten fold increase in capacity. Even if costs went up 100 fold, that doesn't negate an increase in capacity. There's a 100TB SSD available. It's $40,000 but it's AVAILA?BLE even if no one buys it.
A quick search turned up info up to 2017. It seems cost per TB almost stopped dropping around 2017, but there was a big decline 2009 - 2017. The lack of competition obviously is a factor, with only three remaining manufacturers. Even if drives aren't getting cheaper per TB, being able to store more TB per rack saves money for data centers. WD claims increased reliability versus 1 - 10 TB drives, which also reduces TCO. Lastly, HDD manufacturing is not a failing industry. Unit shipment decline is projected to end and turn into an increase soon.
Low capacity HDDs are dead. HDD shipments are cratering. There are fewer and fewer HDDS being sold which means each individual HDD will have to bear a greater proportion of capital, ongoing, and research expense. The silver lining for HDD manufacturers is that the average price per HDD sold is soaring as hyperscalers and data storage warehouses are now the main purchasers of HDDs. This will lead to fewer and fewer new HDD models coming out suitable for home / individual use.
Nicolaim, you said there is a projected increase in unit shipment. That isn't worth the paper it's written on until it actually happens. It might just be a gambit to raise some funding by a storage company. I found an assessment of global HDD shipments 1975-2025, released in May 2020. That rise in HDD shipments is projected to start in 2024, which means it's purely speculative and based on crossed fingers and wild fantasies. https://www.statista.com/statistics/398951/global-...
This is categorically false. HDD shipments are at their highest level ever and still continuing to grow year over year. That you would argue such an incorrect and deliberately false statement indicates your lack of understanding of the market.
Most of the demand is cloud these days and there has big a drop in retail consumer use, but drive shipments remain as high as they've ever been.
That's like the opposite of my observations. The decline stopped in 2011 after the floods in Thailand lead to a very severe shortage. Prices shot straight up at that time, and it took years to get the TB/$ ratio back to 2011 levels, probably not before 2016. Much higher capacities had been introduced in the meantime, but at pricing tiers that just did not exist before.
Looking back through our PO's (ain't QuickBooks great!) I found an order for 2TB WD enterprise drives dated ten years and ten days ago. $249 each. With bulk discounts. A 9x capacity increase for a 2x price increase (given $250 in 2010 ~= $300 today) is not bad.
10 years ago they sold 10x as many non-enterprise drives. The collapse in sales volume and reduced economies of scale has hammered low margin consumer models far harder than enterprise ones.
WD say nothing about exact costs, like whether they have dropped by 10x or whatever. What they say is the new drives offre "unmatched total cost of ownership". Is that a reasonable claim? Why not?
You CAN find cheaper cost per bit for example you can get a (slower...) 12TB drive right now for around $220. But the claim is not cheaper cost per bit, it's TCO.
(a) For COMPARABLE drives, so also enterprise class, cost per bit is close to the same, say around $330 for 12TB. (b) More importantly you have to pay for using these drives in physical space (ie rent), in AC cooling costs, in electricity. Maybe these drives are even more reliable than their predecessors. Without knowing these numbers we have no way to say whether TCO is lower or not. But I'm going to trust WD on this more than some rando on the internet who's upset at something the rest of us cannot see or understand.
You could nitpick that in 2010 the first 3 TB drives were released. But calling this "Fake News" is just wrong. Fake News is what Trump does: to just plain lie and make up things. But ever since humans started to speak, they tended to present information in misleading ways, by leaving things out or implying wrong context. That's not fake, because what is being said is wrong.
So if you want you can accuse AT of misleading the reader by not mentioning the fabrication cost increase. IMO that's still absolutely not justified, because the article simply does not talk about cost. Would they have wanted to mislead there would be chatchy phrases about the cost per GB going down or something like that.
Since these drives are enterprise drives, lets look a enterprise drives from 10 years ago in a 2 TB drive review: https://hothardware.com/reviews/definitive-2tb-har... "All three enterprise drives are the most-expensive ones in the roundup, ranging in price from $290 up to $318." So, yes the price went up from ~$300 to ~$593.
The price per TB went from $150/TB to $33/TB. We get a 9-fold increase in capacity. We get a 4.5-fold decrease in $/TB. We get a doubling of the rated mean time between failures. We get lower idle and lower operating power usage. We get 8-fold increase in cache. We get a 2.5-fold increase in sustained data transfer rate.
Aren't those SMR (shingled) drives? This one here is supposedly CMR with some form of heat assist, but I agree that the price is still a bit steep. I am actually more concerned about the longevity of the drive platters themselves with all that spot heating during writes. That might be just fine for cold storage, but not sure just how the platters themselves would hold up with plenty of writes/rewrites. The refusal of WD to tell us what energy-assisted really means doesn't help to build confidence.
Since NO-ONE -- not anandtech, not WD, claimed that the drives were in any way cheaper than (or even the same price as) the top of the line ten years ago, so what?
We seem to have moved onto the next stage of internet outrage where, if you can't find something to be angry about in the article, get angry at a claim that no-one involved with the article ever made.
But to be fair its not illegal to mismatch real articles alongside tons of ads to make clicking them easier on accident. I mean even this site has the stupid "You won't believe what so and so looks like now!" and other dumb ads.
The statement is essentially correct: in ten years drives are 10x larger in capacity, same size physically. Are you unhappy than 9x is not exactly 10x?
As for cost, that has basically doubled (for that 9x increase in capacity) https://hothardware.com/reviews/definitive-2tb-har... but since no-one made any claims about how costs were changing, again I don't understand why you're so mad.
I'm not going to invest a ton of time here, but a quick internet search shows the Barracuda XT 2TB was priced over $300 when it launched, and that's for a desktop quality drive. Enterprise versions surely would have been more expensive (although perhaps not twice as expensive). I think you're looking for something to complain about that isn't really an issue.
It's interesting to see marketing BS pipeline at work.
"Experts" simply forward the crap that they are fed to. If they had been served some story about schroedinger's cat rearanging the magnetic domains, they'd go happily with that.
With apologies, your comment is a bit nebulous, so I'm not quite sure how to parse it. Just so that I'm clear, what exactly are you saying that we got wrong/incorrect?
While we have science & engineering backgrounds, we don't live and breathe HDD technology like full-on HDD engineers do. So although we aim to get things correct, we're definitely not infallible. In this specific case, we are not aware of any issues with the technology presented by WD or their claims. But if there is something wrong, then please let us know so that we can correct it in this article and ensure it remains correct going forward.
His comment is not just nebulous, it is nothing less than a Schroedingers comment. A comment that is in a superposition of saying everything and nothing at once. I mean, come on, who would be so cruel to put a comment together with a bottle of poisonous gas in a sealed dark box? Comment abuse is what it is...
One thing I've never been able to figure out is why, now that each head has its own micro-actuator, no one has made a hard drive that can read or write more than one platter at the same time. That wouldn't do much of anything to improve performance for random reads and writes, but would allow sequential I/O operations to be much faster. For example, if a single head can only read 150 MB/sec, but you have 6 heads, sequential read performance should be 900 MB/sec. If a 16TB disk in a RAID array had to be replaced, the replacement disk could be filled with data in less than half an hour. (OK, more if you are limited by the bandwidth to the host, but still, you get the point. Sequential performance matters for RAID array rebuilds even if your workload is purely random.)
Maybe the drive manufacturers have been using essentially the same controller technology for years and it would cost too much to come up with a new design that could access multiple heads at once?
I'm pretty sure the answer to this is that it would be almost impossible to guarantee the ability to write to a continuous track of sufficient length across all platters. You'd run a tremendous risk of overwriting data in the middle of a write and/or have wildly fluctuating I/O rates, not to mention that it would dramatically increase controller complexity.
I suspect this hasn't happened for much the same reason that SSHDDs died a dismal death: it's cheaper just to solve the speed problem with an SSD.
The prices on HDDs haven't followed the previous trends... Yes, we're getting more capacity, but prices are higher.
Looking at my amazon order history, I was buying 4 TB HDDs for 139.99 back in 2013. If prices were following the previous trends, we should have 32 TB HDDs at 140 bucks right now. We're very far off from those prices and capacities...
The upside is, SSDs do seem to be on an exponential trend for price perf latley.
The article was informative for me. I had never seen a three stage actuator, which is amazing from a mechanics standpoint. It's interesting to see the world of physics being canvased for any and all tech to increase HDD capacity.
Sorry to see the inflammatory comments on this post. Seems like cost & price are being endlessly debated in other articles as well, sometimes over as little as $10/$20. Best to focus on the tech specifications and performance, and leave the price to those who are actually buying & selling these things.
Just checked my previous purchases: WD Red 4TB Oct 2013 £155 £38.75/TB bought WD Red 4TB Feb 2014 £140 £35/TB bought WD Red 10TB Mar 2019 £296 £29.60/TB bought (first one was DOA, replacement be fine, 24/7 in a NAS) WD Red 14TB Jul 2020 £393 £28.07/TB considering. I will maybe hold out for 16/18TB Reds, this news may suggest they could be be out soon. I just learnt that there is a now two 10TB Reds. 100EFAX VS 101EFAX, internet says the 101 is air filled so runs with slightly more power, heat and noise, but looks to be cheaper.
"WD's first-generation EAMR technology (christened as ePMR) involves the application of electrical current to the writing head's main pole (this is in addition to the current sent through the voice coil) during write operations. " -- In conventional PMR writing, current is applied to the main pole. In ePMR as shown in the graphic provided, additional current applied to the shields to enhance write performance. The voice coil is used to position the head on the correct track during writing and reading by varying the current applied to it.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
69 Comments
Back to Article
azfacea - Thursday, July 9, 2020 - link
phake news."Around 10 years back, we had 2TB 3.5" HDDs ... In the last decade, we have seen advancements in three different categories that have enabled a 10-fold increase in the capacity of HDDs while retaining the same physical footprint"
its not anywhere near the same cost, your 10 fold increase is propaganda and misinformation. nevermind that your cost numbers are not TCO. I wouldnt be surprised if its a sponsored article by a failing industry (not saying it is)
Gigaplex - Thursday, July 9, 2020 - link
Nowhere in that quote mentioned anything about cost. It's a 10-fold increase in capacity.azfacea - Thursday, July 9, 2020 - link
thats how phake news works. selective and misleading misinformationmelgross - Thursday, July 9, 2020 - link
You don’t know what you’re talking about. Just look at capacities and their prices. Even a child can do that. When I see a 2TB drive that used to cost $600, and now see an 18TB drive that costs $600, I’m pretty impressed, and you should be too.It’s getting more difficult to increase capacities as time goes on. The sophisticated technology used these days is much more complex. It’s pretty surprising that it can even be done at these prices. I remember the days when the first had for personal computers came out, many years ago. It was 3.125MB, and cost $3,125. In today’s dollars, that’s almost $6,000. There’s been steady progress ever since.
ZoZo - Thursday, July 9, 2020 - link
"When I see a 2TB drive that used to cost $600"What?? 2TB drives certainly never cost anywhere close to $600.
The first one launched in january 2009 with an MSRP of $250, but the actual price in stores nearly immediately dipped well below $200.
In july 2010, which is 10 years ago, they cost around $100, 1/6th of what you're quoting.
During the summer of 2011, before the dreadful floods of fall 2011 in Thailand, they cost below $60, less than 1/10th of what you're quoting.
close - Friday, July 10, 2020 - link
Oh haven't you learned by now to ignore azfacea and their ramblings? They're like those cooks with "the end is nigh" on every corner if anyone will ever listen.@azfacea, don't worry, hard disks are fake news, they don't actually exist. They stopped existing the moment you started rambling how they'll be completely replace by SSD is a matter of months. Now we're just buying the for collections. ;)
close - Friday, July 10, 2020 - link
This was supposed to be a reply for @melgross. Otherwise indeed, even the enterprise 2TB WDs were ~$300-350 at launch.Spunjji - Friday, August 28, 2020 - link
It seems that a lot of people have forgotten what hard drive prices were like before 2011. They've never really recovered from that.Hrunga_Zmuda - Monday, November 29, 2021 - link
Dude, I remember when I was buying a 1GB drive for $330 (I'm old, get over it.) the dealer went into the back of the room and brought out the biggest drive he had ever sold. It was 10TB, and it cost $10,000.At the same time, at the newspaper I worked for, we had a pagination system (design the newspaper on computer) that ran on a Sun Workstation (the old pizza box models). It had a 1 gigabyte storage system that was larger than a washing machine and cost $100,000.
You clearly need to listen more.
Hrunga_Zmuda - Monday, November 29, 2021 - link
Sorry, that $10,000 drive was 10 GB, not TB.burnte - Thursday, July 9, 2020 - link
Actually, fake news is about sources that completely make up information, such as the moon landing was staged, or JFK isn't dead, etc. When it came up in 2016 it was about sites that would publish articles that Obama was going to institute martial law rather than leave, or that the Pope endorsed Trump, or that RuPaul accused Trump of something. All of those were fake news.Inaccurate stories is something else. However, stating that we've seen a tenfold increase in capacity is either true or it's not. Not mentioning cost or other facts has no bearing on if there was a ten fold increase in capacity. Even if costs went up 100 fold, that doesn't negate an increase in capacity. There's a 100TB SSD available. It's $40,000 but it's AVAILA?BLE even if no one buys it.
FunBunny2 - Saturday, July 11, 2020 - link
"There's a 100TB SSD available. It's $40,000 but it's AVAILA?BLE"I remember STEC was the vanguard of high capacity, high cost (SLC) SSD ages ago. Who's making this sort?
Santoval - Friday, July 10, 2020 - link
You are trolling now aren't you?sonicmerlin - Saturday, July 11, 2020 - link
Can you stop polluting the comments section with your childish nonsense?StevoLincolnite - Thursday, July 9, 2020 - link
His point was that it came at a corresponding increase in price.nicolaim - Thursday, July 9, 2020 - link
A quick search turned up info up to 2017. It seems cost per TB almost stopped dropping around 2017, but there was a big decline 2009 - 2017.The lack of competition obviously is a factor, with only three remaining manufacturers.
Even if drives aren't getting cheaper per TB, being able to store more TB per rack saves money for data centers. WD claims increased reliability versus 1 - 10 TB drives, which also reduces TCO.
Lastly, HDD manufacturing is not a failing industry. Unit shipment decline is projected to end and turn into an increase soon.
Tomatotech - Thursday, July 9, 2020 - link
Low capacity HDDs are dead. HDD shipments are cratering. There are fewer and fewer HDDS being sold which means each individual HDD will have to bear a greater proportion of capital, ongoing, and research expense. The silver lining for HDD manufacturers is that the average price per HDD sold is soaring as hyperscalers and data storage warehouses are now the main purchasers of HDDs. This will lead to fewer and fewer new HDD models coming out suitable for home / individual use.Nicolaim, you said there is a projected increase in unit shipment. That isn't worth the paper it's written on until it actually happens. It might just be a gambit to raise some funding by a storage company. I found an assessment of global HDD shipments 1975-2025, released in May 2020. That rise in HDD shipments is projected to start in 2024, which means it's purely speculative and based on crossed fingers and wild fantasies. https://www.statista.com/statistics/398951/global-...
vortmax2 - Thursday, July 9, 2020 - link
"Nicolaim, you said there is a projected increase in unit shipment. That isn't worth the paper it's written on until it actually happens.""Unit shipment decline is projected to end and turn into an increase soon."
That's why they called a "projection" or in the finance industry an outlook, not an actual.
rahvin - Friday, July 10, 2020 - link
<HDD shipments are cratering.>This is categorically false. HDD shipments are at their highest level ever and still continuing to grow year over year. That you would argue such an incorrect and deliberately false statement indicates your lack of understanding of the market.
Most of the demand is cloud these days and there has big a drop in retail consumer use, but drive shipments remain as high as they've ever been.
s.yu - Saturday, July 11, 2020 - link
Then if anything $/GB should continue to fall, unlike this reality.FunBunny2 - Saturday, July 11, 2020 - link
"each individual HDD will have to bear a greater proportion of capital, ongoing, and research expense. "the Tyranny of Average Cost at work. same reason a Ferrari cost more than Yugo.
ZoZo - Thursday, July 9, 2020 - link
That's like the opposite of my observations. The decline stopped in 2011 after the floods in Thailand lead to a very severe shortage. Prices shot straight up at that time, and it took years to get the TB/$ ratio back to 2011 levels, probably not before 2016. Much higher capacities had been introduced in the meantime, but at pricing tiers that just did not exist before.Kvaern1 - Saturday, July 11, 2020 - link
"A quick search turned up info up to 2017. It seems cost per TB almost stopped dropping around 2017, but there was a big decline 2009 - 2017."2017 was the year of the Thailand flood. Apparently that was good for business.
ksgtjxBAbK - Thursday, July 9, 2020 - link
Looking back through our PO's (ain't QuickBooks great!) I found an order for 2TB WD enterprise drives dated ten years and ten days ago. $249 each. With bulk discounts. A 9x capacity increase for a 2x price increase (given $250 in 2010 ~= $300 today) is not bad.ZoZo - Thursday, July 9, 2020 - link
10 years ago, non-enterprise ones cost $100 in online stores.DanNeely - Friday, July 10, 2020 - link
10 years ago they sold 10x as many non-enterprise drives. The collapse in sales volume and reduced economies of scale has hammered low margin consumer models far harder than enterprise ones.s.yu - Saturday, July 11, 2020 - link
Then their margins are higher, at the cost of consumers of course.crimsonson - Thursday, July 9, 2020 - link
Yea - the OP is making comments on something that was not mentioned in the quote. Puzzling.MrVibrato - Thursday, July 9, 2020 - link
I guess azfacea refers to the actual WD press release (linked at the end of the article), which talks about TCO...name99 - Thursday, July 9, 2020 - link
WD say nothing about exact costs, like whether they have dropped by 10x or whatever.What they say is the new drives offre "unmatched total cost of ownership".
Is that a reasonable claim? Why not?
You CAN find cheaper cost per bit for example you can get a (slower...) 12TB drive right now for around $220. But the claim is not cheaper cost per bit, it's TCO.
(a) For COMPARABLE drives, so also enterprise class, cost per bit is close to the same, say around $330 for 12TB.
(b) More importantly you have to pay for using these drives in physical space (ie rent), in AC cooling costs, in electricity. Maybe these drives are even more reliable than their predecessors.
Without knowing these numbers we have no way to say whether TCO is lower or not.
But I'm going to trust WD on this more than some rando on the internet who's upset at something the rest of us cannot see or understand.
Deicidium369 - Thursday, July 9, 2020 - link
WD Gold WD181KRYZ 18TB 7200 RPM 512MB Cache SATA 6.0Gb/s 3.5" Internal Hard Drive $593 - the drive listed in the article. Price from NewEgg.DigitalFreak - Friday, July 10, 2020 - link
Nah, he just likes to throw around the latest buzzphrase to act like he's cool or something.MrSpadge - Thursday, July 9, 2020 - link
You could nitpick that in 2010 the first 3 TB drives were released. But calling this "Fake News" is just wrong. Fake News is what Trump does: to just plain lie and make up things. But ever since humans started to speak, they tended to present information in misleading ways, by leaving things out or implying wrong context. That's not fake, because what is being said is wrong.So if you want you can accuse AT of misleading the reader by not mentioning the fabrication cost increase. IMO that's still absolutely not justified, because the article simply does not talk about cost. Would they have wanted to mislead there would be chatchy phrases about the cost per GB going down or something like that.
dullard - Thursday, July 9, 2020 - link
Since these drives are enterprise drives, lets look a enterprise drives from 10 years ago in a 2 TB drive review:https://hothardware.com/reviews/definitive-2tb-har...
"All three enterprise drives are the most-expensive ones in the roundup, ranging in price from $290 up to $318."
So, yes the price went up from ~$300 to ~$593.
The price per TB went from $150/TB to $33/TB. We get a 9-fold increase in capacity. We get a 4.5-fold decrease in $/TB. We get a doubling of the rated mean time between failures. We get lower idle and lower operating power usage. We get 8-fold increase in cache. We get a 2.5-fold increase in sustained data transfer rate.
vFunct - Thursday, July 9, 2020 - link
So, a little bit more than inflation.nicolaim - Thursday, July 9, 2020 - link
It's way more than inflation, but as npz mentions below, Seagate 16 TB drives retail for $382.eastcoast_pete - Thursday, July 9, 2020 - link
Aren't those SMR (shingled) drives? This one here is supposedly CMR with some form of heat assist, but I agree that the price is still a bit steep. I am actually more concerned about the longevity of the drive platters themselves with all that spot heating during writes. That might be just fine for cold storage, but not sure just how the platters themselves would hold up with plenty of writes/rewrites. The refusal of WD to tell us what energy-assisted really means doesn't help to build confidence.cjl - Thursday, July 9, 2020 - link
No, at least the Exos and IronWolf 16TB drives are all CMR, not SMR.eastcoast_pete - Saturday, July 11, 2020 - link
If so, I'll give them another look. Could use a couple of 16 TB CMR drives.name99 - Thursday, July 9, 2020 - link
Since NO-ONE -- not anandtech, not WD, claimed that the drives were in any way cheaper than (or even the same price as) the top of the line ten years ago, so what?We seem to have moved onto the next stage of internet outrage where, if you can't find something to be angry about in the article, get angry at a claim that no-one involved with the article ever made.
prisonerX - Thursday, July 9, 2020 - link
Straw man arguments are the stock in trade of these idiots. Comment sections are classic example of the tragedy of the commons.Ryan Smith - Thursday, July 9, 2020 - link
"I wouldnt be surprised if its a sponsored article by a failing industry (not saying it is)"And it isn't. If it were a sponsored article, it would be very clearly labeled as such. It's quite illegal not to disclose such things.
imaheadcase - Thursday, July 9, 2020 - link
But to be fair its not illegal to mismatch real articles alongside tons of ads to make clicking them easier on accident. I mean even this site has the stupid "You won't believe what so and so looks like now!" and other dumb ads.prisonerX - Thursday, July 9, 2020 - link
Just install an ad blocker and stop whining.name99 - Thursday, July 9, 2020 - link
What's your complaint? Are you saying we had larger than 2TB drives in 2010?3TB was introduced at the end of 2010 (technically later than "ten years ago")
https://www.anandtech.com/show/3858/the-worlds-fir...
Till then the hot excitement, the state of the art, for 2010, was 2TB
https://www.anandtech.com/show/2866
The statement is essentially correct: in ten years drives are 10x larger in capacity, same size physically. Are you unhappy than 9x is not exactly 10x?
As for cost, that has basically doubled (for that 9x increase in capacity)
https://hothardware.com/reviews/definitive-2tb-har...
but since no-one made any claims about how costs were changing, again I don't understand why you're so mad.
abhaxus - Thursday, July 9, 2020 - link
I'm not going to invest a ton of time here, but a quick internet search shows the Barracuda XT 2TB was priced over $300 when it launched, and that's for a desktop quality drive. Enterprise versions surely would have been more expensive (although perhaps not twice as expensive). I think you're looking for something to complain about that isn't really an issue.mukiex - Monday, July 13, 2020 - link
"Around 5 years ago, we had 8TB 3.5" HDDs for $150. Today, they cost the same fucking price to the same."So much progress.
Brane2 - Thursday, July 9, 2020 - link
It's interesting to see marketing BS pipeline at work."Experts" simply forward the crap that they are fed to.
If they had been served some story about schroedinger's cat rearanging the magnetic domains, they'd go happily with that.
Whatever sells adverts.
MrSpadge - Thursday, July 9, 2020 - link
Care to educate us on the truth?prisonerX - Thursday, July 9, 2020 - link
Yes clearly experts are the problem, not addled and ignorant anonymous posters in internet forums bearing conspiracy theories.Ryan Smith - Thursday, July 9, 2020 - link
Hi Brane2,With apologies, your comment is a bit nebulous, so I'm not quite sure how to parse it. Just so that I'm clear, what exactly are you saying that we got wrong/incorrect?
While we have science & engineering backgrounds, we don't live and breathe HDD technology like full-on HDD engineers do. So although we aim to get things correct, we're definitely not infallible. In this specific case, we are not aware of any issues with the technology presented by WD or their claims. But if there is something wrong, then please let us know so that we can correct it in this article and ensure it remains correct going forward.
-Thanks
Ryan Smith
sheh - Friday, July 10, 2020 - link
His comment is nebulous more than a bit. Not worth your reply time.MrVibrato - Saturday, July 11, 2020 - link
His comment is not just nebulous, it is nothing less than a Schroedingers comment. A comment that is in a superposition of saying everything and nothing at once. I mean, come on, who would be so cruel to put a comment together with a bottle of poisonous gas in a sealed dark box? Comment abuse is what it is...KAlmquist - Friday, July 10, 2020 - link
One thing I've never been able to figure out is why, now that each head has its own micro-actuator, no one has made a hard drive that can read or write more than one platter at the same time. That wouldn't do much of anything to improve performance for random reads and writes, but would allow sequential I/O operations to be much faster. For example, if a single head can only read 150 MB/sec, but you have 6 heads, sequential read performance should be 900 MB/sec. If a 16TB disk in a RAID array had to be replaced, the replacement disk could be filled with data in less than half an hour. (OK, more if you are limited by the bandwidth to the host, but still, you get the point. Sequential performance matters for RAID array rebuilds even if your workload is purely random.)Maybe the drive manufacturers have been using essentially the same controller technology for years and it would cost too much to come up with a new design that could access multiple heads at once?
sheh - Friday, July 10, 2020 - link
Maybe you need a larger movement range.Or maybe it's being worked on and would finally materialize in a few years. One can hope.
DanNeely - Friday, July 10, 2020 - link
Could be that two heads on the same arm trying to write at once would interfere with each other.extide - Friday, July 10, 2020 - link
That is how they work. On a 9 platter drive all 18 heads are reading/writing at the same time, the data is essentially striped across them.Kvaern1 - Saturday, July 11, 2020 - link
This, and it's noticeably the reason why larger HDs are faster than smaller ones from the same series.sheh - Sunday, July 12, 2020 - link
There are mainly (or only?) two things that increase linear speed: RPM, and linear data density.Not directly related, but here's an interesting study on HDD geometry:
http://blog.stuffedcow.net/2019/09/hard-disk-geome...
sheh - Sunday, July 12, 2020 - link
That's not how they work. If it were, A 9 platter drive would be 9 times faster than a 1 platter drive.Spunjji - Friday, August 28, 2020 - link
I'm pretty sure the answer to this is that it would be almost impossible to guarantee the ability to write to a continuous track of sufficient length across all platters. You'd run a tremendous risk of overwriting data in the middle of a write and/or have wildly fluctuating I/O rates, not to mention that it would dramatically increase controller complexity.I suspect this hasn't happened for much the same reason that SSHDDs died a dismal death: it's cheaper just to solve the speed problem with an SSD.
Kamus - Friday, July 10, 2020 - link
The prices on HDDs haven't followed the previous trends... Yes, we're getting more capacity, but prices are higher.Looking at my amazon order history, I was buying 4 TB HDDs for 139.99 back in 2013. If prices were following the previous trends, we should have 32 TB HDDs at 140 bucks right now. We're very far off from those prices and capacities...
The upside is, SSDs do seem to be on an exponential trend for price perf latley.
voicequal - Saturday, July 11, 2020 - link
The article was informative for me. I had never seen a three stage actuator, which is amazing from a mechanics standpoint. It's interesting to see the world of physics being canvased for any and all tech to increase HDD capacity.Sorry to see the inflammatory comments on this post. Seems like cost & price are being endlessly debated in other articles as well, sometimes over as little as $10/$20. Best to focus on the tech specifications and performance, and leave the price to those who are actually buying & selling these things.
keyserr - Sunday, July 12, 2020 - link
Just checked my previous purchases:WD Red 4TB Oct 2013 £155 £38.75/TB bought
WD Red 4TB Feb 2014 £140 £35/TB bought
WD Red 10TB Mar 2019 £296 £29.60/TB bought (first one was DOA, replacement be fine, 24/7 in a NAS)
WD Red 14TB Jul 2020 £393 £28.07/TB considering. I will maybe hold out for 16/18TB Reds, this news may suggest they could be be out soon.
I just learnt that there is a now two 10TB Reds. 100EFAX VS 101EFAX, internet says the 101 is air filled so runs with slightly more power, heat and noise, but looks to be cheaper.
breynded - Monday, July 13, 2020 - link
"WD's first-generation EAMR technology (christened as ePMR) involves the application of electrical current to the writing head's main pole (this is in addition to the current sent through the voice coil) during write operations. "-- In conventional PMR writing, current is applied to the main pole. In ePMR as shown in the graphic provided, additional current applied to the shields to enhance write performance. The voice coil is used to position the head on the correct track during writing and reading by varying the current applied to it.