60TB Hard Drives within the decade

60TB Hard Drives within the decade

The next advance, which will take disk drives to 5Tbit/in areal density, is heat-assisted magnetic recording (HAMR). HAMR, which Seagate patented in 2006, adds a small laser to a drive to change the magnetic properties of the disk.

HAMR also uses nanotube-based lubrication to allow the read/write head of a disk drive to get closer to the surface of a spinning platter to write and read more bits of data.

A chief advance with HAMR is the switch from a cobalt platinum alloy, the coating used on today's disks for data bit recording, to iron platinum, a much stronger magnetic material that helps stabilize data bits at smaller sizes.

Seagate's first HAMR drives are expected in 2015 or 2016.

http://www.computerworld.com/s/article/9227382/60TB_disk_drives_could_be_a_reality_in_2016

The article is from 2012... 4TB harddisks were just available, now 2 years down the road, the first 6TB disks can be bought but are premium... i don't think in 2 years time those 6TB will be replaced by 60TB. Unless some hi-tecsuperduperlargehydroncollidertechiebreakthrough happened. Which is possible

I dont think hdd manufacturers will release extremely large drives very quickly becase they will harm their own business.  They will use the technology to enable larger drives but will do it sowly and gradually.

Unless some hi-tecsuperduperlargehydroncollidertechiebreakthrough happened.

That's the point... it has happened.  HAMR with iron platinum increases storage density dramatically and HAMR based drives are just a year or 2 away.  This is analogous to the switch from linear to perpendicular recording.

http://www.computerworld.com/slideshow/detail/145923#slide1

"Data storage is among the few techological advances that has actually surpassed our current needs."

"Data storage is among the few techological advances that has actually surpassed our current needs."

"Needs" have a way of growing quickly to exceed whatever the current available capacity is 

I remember when I bought my first hard drive [26MB for $4500] that it seemed like SO much storage it would probably last for YEARS.    Two years later it was full; and a couple years after that I simply threw it away !!

I remember when I bought my first hard drive [26MB for $4500] that it seemed like SO much storage it would probably last for YEARS.    Two years later it was full; and a couple years after that I simply threw it away !!

Yep I remember those days. I got a 105MB drive in my first computer. I got a lot of ribbing because it was a "senseless amount of storage" and that "I'll never fill it". 25 years later I have 250,000+ times that storage and that might get me to the end of the year. Kinda crazy when you think about it.

I dont think hdd manufacturers will release extremely large drives very quickly becase they will harm their own business.  They will use the technology to enable larger drives but will do it sowly and gradually.

I seriously doubt manufacturers will hold back the tech if it's market ready. They will most likely just charge an exuberant amount for it to maximize the profits in both market segments. Just a guess...

Resurrecting this thread rather than posting a new one.

Looks like reported timelines have been pushed back to 2017, which probably means 2018.

Still time to wait until until 100TB drives 

Plus, the companies are making a pretty penny on 4-8TB drives still. Why rush a massive TB capacity drive.

http://www.kitguru.net/components/hard-drives/anton-shilov/seagate-demos-hamr-hdds-vows-to-start-commercial-shipments-in-late-2017/

Still time to wait until until 100TB drives 

Geez just imagine the parity check time (assuming the speeds remains similar to the drives today)...

Still time to wait until until 100TB drives 

 

Geez just imagine the parity check time (assuming the speeds remains similar to the drives today)...

I was just sat sipping my morning coffee thinking the same thing!

I think the traditional method of doing that is going to have to change (if it can) to cope with drives of increasing capacity.

Can you really wait over a month for a Parity Check? He he! Quite funny!!

What size would you buy if it were available? Let's say cost was $40/T regardless of size. So a 60T drive would cost $2400.

What drive sizes would you be buying?

What size would you buy if it were available? Let's say cost was $40/T regardless of size. So a 60T drive would cost $2400.

 

What drive sizes would you be buying?

I'm making an assumption of pricing trends based on History. I'm assuming that as the capacity of drives increase the cost per TB will decrease. Therefore I expect these "high capacity" drives will be more affordable than you're suggesting. I am guessing though.

However, to answer your question. Assuming above, with a 100TB parity protected drive I could have a redundant 2 drive 100TB system in the "smallest" of form factors. 2 drives in each system (Backup and Main) and 2 drives off site. Rather than the 12 I have now. Reduce number of drives by half and have 75% more space than I do today. No planned upgrades for at least 5 years based on my usage and increasing requirements.

Given the technology is coming, I feel the practicalities of the "administration" associated with the use of such large capacity drives is what will make their use pointless. If you have a setup as I mention above but the software that uses them is limited to being able to do things like Parity Checking and Rebuilding over timelines such as a Month +. Not practical.

What size would you buy if it were available? Let's say cost was $40/T regardless of size. So a 60T drive would cost $2400.

 

What drive sizes would you be buying?

The ones for under $30/T.

Bigger SSDs are becoming available, with Samsung selling 2TB units.  Bigger drives on their way.  HDDs are going to have to get bigger to compete.

4K video uses a lot of space, and it's becoming more and more popular.

As I noted earlier, "Needs" have a way of growing quickly to exceed whatever the current available capacity is 

Drive capacities will indeed continue to grow ... and we will undoubtedly continue to fill them and wish they were even larger !!    Drives that seem "small" today (2TB or 3TB units) were considered massive not all that many months ago.

60 TB drive in a windows desktop could be the end of unraid .....

no need for extra hardware .... all depends on the cost of that drive ....

second one for backup and you have an array of 2 drives instead of 2 x 11 drives (if you count 6tb drives which are available now)+ 2 x sets of hardware

less electricity...

less heat in my office

think i am going to buy a few more second hand 2tb drives to replace the 1tb 's is still have and hope these 60tb drive will not be too expensive

was considering buying 2 x 4 TB drives for parity drives in my 2 servers and gradually replace when break .... but these super big drives may postpone/anullate any big changes that i was thinking off....

I wouldn't hold my breath for these 60T drives showing up any time soon. And when they do the price is not going to be cheap / TB for a very very long time.

Not sure if others have noticed, but prices for large drives have been going up not down. 6T have been out for a very long time and their $/T are not the lowest. In fact 3T are still cheaper/T than 4T drives. 3 years ago, new larger drives would be cheaper per TB within 3-4 months.

I was buying HGST 4T drives in the $130s and $140s a year ago. Now you'll spend $160s or $170s.

The times they are a changing.

And I would never want a 60TB parity. My 6T parity is only costing about $250, and with 10 data drives can support a 60TB array. A 60TB drive is going to be in the thousands. A bad sector or two is going to be a much bigger deal one one of these monsters! You'll be crying in your Cheerios!

I wouldn't be so sure that 60TB drives will be "in the thousands".    The last hard drive I bought that cost in the thousands was a 26MB Seagate ~ 1980 that cost me $4500.    Today's 8TB drives are 307,692 times as large and cost 1/15th as much => in other words 1/4,615,380th as much per unit of storage.

While I certainly don't anticipate that degree of price differential in the next few evolutions in the hard drive market; I also don't think the larger drives will cross the $1,000/drive mark, regardless of capacity ... certainly not the consumer versions anyway (high-end enterprise units may be a different story).

Put another way, you can buy 4 8TB drives for ~ $1000 today => that's 32TB of storage.    I don't think a 32TB drive will cost more than that, and it will likely be notably less.    And 60TB isn't all that much of an evolution from that point.

I don't know about 60TB in the relatively near future, but I certainly wouldn't be surprised to see 16TB, or even 32TB within just a couple years.

Can you really wait over a month for a Parity Check? He he! Quite funny!!

But you likely wouldn't. Since the HAMR technology achieves the massive capacity increases due to much higher areal density, read/write speeds would also increase accordingly due to the increased areal density, all other things being equal (which may or may not be the case).

So assuming all other things being equal, with a roughly 10x increase in areal density you could expect a roughly 10x increase in read/write speed at the same RPM. That said, transferring that data to the bus today would be limited by SATA speeds as SATA III is roughly only 4x faster than current gen spinners. A PCIe interface (or some other new interface) would be needed to get anywhere near what one of these future 60TB drives could likely like deliver as far as transfer speeds. Point being, it's entirely possible that one of these not-so-far-in-the-future 60TB drives could have higher throughput rates than current SSDs.

Agree => dramatically higher capacities will almost certainly mean much higher areal densities ... with corresponding increases in sustained transfer rates.    While as long as drives are spinning they're not going to match SSD access times; they could easily achieve SSD-level sustained transfer rates (or even higher).

And I'd certainly expect interface technology to maintain the ability to handle the higher areal density transfer rates.  Note that 12Gb SAS interfaces are already available ... drives interfacing at this speed could have roughly 6 times the areal density of today's 1TB/platter drives and still not be bottlenecked by the interface.    And I suspect both SATA and SAS haven't yet seen the end of their evolution 

I wouldn't be so sure that 60TB drives will be "in the thousands".    The last hard drive I bought that cost in the thousands was a 26MB Seagate ~ 1980 that cost me $4500.    Today's 8TB drives are 307,692 times as large and cost 1/15th as much => in other words 1/4,615,380th as much per unit of storage.

In the 80s and 90s we saw modem speeds increase from 1200 baud to 19.2k baud and more with compression. But as we reached limits, we saw a halt in speed increases - and a move to network vs phone based telecommunications. We saw rapid increases in network speed from 10 to 100 to 1000 Mb/sec and pretty much a halt for 10 years. We have seen processor speeds that used to double with a new generation CPU, being reduced to a few percentage points - and it is increasing parallel performance not raw sequential performance. Graphics processing increases have slowed too, facing heat and frequency limits. In short, as we approach some practical and theoretical limits, progress slows. Drives have similarly ridden the technology curve and seen dramatic improvements in density and platters, but that has been slowing of late as limits are reached. Will HAMR allow 60T drives that run circles around the current drives in speed and, at the same time, cut the $/T to 1/3 of today - let's just say I am skeptical.

While I certainly don't anticipate that degree of price differential in the next few evolutions in the hard drive market; I also don't think the larger drives will cross the $1,000/drive mark, regardless of capacity ... certainly not the consumer versions anyway (high-end enterprise units may be a different story).

Are individuals, besides us OCD NAS users, going to be buying 60TB drives? I think not. And if we did - we'd be buying a small handful. Three of them would produce arrays larger than any in existence today. The real market here is the cloud. If the $/T are similar to today's drives, they will buy these monsters for a 10x reduction in the number of "pods" needed for expansion. Drive manufacturers don't have to drive prices to the floor to create demand for these. On the low capacity end, spinners are loosing. In the laptop and workstation market, they will continue to loose market share to SSDs, and soon will be the exception.

Put another way, you can buy 4 8TB drives for ~ $1000 today => that's 32TB of storage.    I don't think a 32TB drive will cost more than that, and it will likely be notably less.    And 60TB isn't all that much of an evolution from that point.

You are comparing archive media with high speed media. The fastest spinners are 6TB (except helium which is MUCH more expensive), and they cost in the $300 price range. Maintaining a similar $/T, a 60TB drive would cost $3000. Will Seagate or WD be saying - ok, we'll cut the $/T to a third of current drives and make up for it in high volume sales?  There is a lot of R&D that has gone into HAMR, and if it works, these drives will demand a premium for some significant time to come, unless ... (see below)

I don't know about 60TB in the relatively near future, but I certainly wouldn't be surprised to see 16TB, or even 32TB within just a couple years.

I agree that drive sizes are not going up by 8x overnight. We have seen drives go from ...

250G -> 320G -> 500G -> 750G -> 1T -> 1.5T -> 2T -> 3T -> 4T -> 6T -> 8T.

(I skipped 5T as it was not a popular size and seemed to happen in parallel with 6T).

I think we'll see 10T then 12T, then maybe 16T. Next couple of years? We'll see. unRAID media moguls may see 8T as but a eyedropper for their needs, the masses see even 4T as a monster size to never be filled. Bigger drives, by their vary nature, REDUCE unit demand. If the spinners are going to survive, they need to increase unit profit. That means getting bigger and cheaper to make, increasing profit per unit.

The best chance of significant advancement in HD size, speed, and price reductions is SSD technology pushing. SSD is faster but also smaller capacity and more expensive. But we know if prices are about the same, SSDs will win every day of the week. And SSD technology is in its infancy compared to hard drives - and innovation is still occurring at a fast pace. We're already seeing SSDs at 1T, and I expect to see them continue to grow in size and reduce in price/TB. If the drive manufacturers want to stay ahead of that, they may put down the HAMR and we might see something dramaticly bigger. Not holding my breath - but that's our best chance IMO. But it won't be a consumer drive or anywhere near cheap IMO.

You have to remember 10TB drives are already a reality. Oh, you want to exclude them because they are not consumer drives? Well, what do you think the 60TB will be?

$/TB for all storage will continue to drop because without a lower $/TB the drives wont be made. 14EB at $19/TB does not care what size the drives are. You want to sell storage, find a way to lower $/TB.

You have to remember 10TB drives are already a reality. Oh, you want to exclude them because they are not consumer drives? Well, what do you think the 60TB will be?

 

$/TB for all storage will continue to drop because without a lower $/TB the drives wont be made. 14EB at $19/TB does not care what size the drives are. You want to sell storage, find a way to lower $/TB.

I don't agree. 6TB drives are more expensive per TB vs 4TB, and 4TB are more expensive than 3TB drives. So why would anyone buy the larger drives? It means fewer drives, fewer pods, less physical space, less support effort, etc. I believe that buyers will continue to pay a premium for larger drives. I certainly think the BackBlazes of the world would look hard at higher capacity drives even if the price is higher if, in the larger equation, costs are reduced. Building only 10% as many pods certainally has a significant cost savings!

You have to remember 10TB drives are already a reality. Oh, you want to exclude them because they are not consumer drives? Well, what do you think the 60TB will be?

 

$/TB for all storage will continue to drop because without a lower $/TB the drives wont be made. 14EB at $19/TB does not care what size the drives are. You want to sell storage, find a way to lower $/TB.

 

I don't agree. 6TB drives are more expensive per TB vs 4TB, and 4TB are more expensive than 3TB drives. So why would anyone buy the larger drives? It means fewer drives, fewer pods, less physical space, less support effort, etc. I believe that buyers will continue to pay a premium for larger drives. I certainly think the BackBlazes of the world would look hard at higher capacity drives even if the price is higher if, in the larger equation, costs are reduced. Building only 10% as many pods certainally has a significant cost savings!

$/TB includes all the racks, sqft, cooling, CPU, memory, cabling, etc. Not just the bare drive pricing. When buying truckloads of drives, the lower $/TB determines which size drive gets purchased.

$/TB includes all the racks, sqft, cooling, CPU, memory, cabling, etc. Not just the bare drive pricing. When buying truckloads of drives, the lower $/TB determines which size drive gets purchased.

$/TB is usually something quoted for drives themselves. Otherwise it is not something we can objectively compare.

But I agree - each organization has to weigh all these types of variables for their unique operations to make good business decisions. As stated, 60 TB drives would offer a lot of cost saving advantages to organizations that go through drives like water, even if the drives themselves are at a premium.