60TB Hard Drives Arriving in 2028, According To Industry Roadmap (tomshardware.com) 43
An anonymous reader shares a report: The arrival of energy-assisted magnetic recording (EAMR) technologies like Seagate's HAMR will play a crucial role in accelerating HDD capacity growth in the coming years. According to the new IEEE International Roadmap for Devices and Systems Mass Data Storage, we will see 60 TB hard disk drives in 2028. If the prediction is accurate, we will see HDD storage capacity doubling in just four years, something that did not happen for a while. Also, IEEE believes that HDD unit sales will increase.
IEEE's latest HDD development roadmap spans 2022 to 2037 and covers 15 years of hard drive evolution. The arrival of HAMR in 2024 will play a pivotal role in the increase in HDD capacity (even though Western Digital has managed to stay competitive with Seagate's HAMR HDDs using a set of its technologies) over the next few years. IEEE engineers expect HDDs to leapfrog to 40TB in 2025 and 60TB in 2028, doubling capacity from 30TB in 2024. By 2037, there will be 100TB of storage space, according to IEEE.
To get to those extreme capacities, HDD makers will have to increase the areal density of their platters steadily. To get to 40TB per drive, they will have to get to 2 TB/inch^2 in 2025 and then to over 4 TB/inch^2 in 2028 to build 60TB HDDs. By 2037, areal density will grow to over 10 Tb/inch^2. Increasing areal density will necessitate the use of new media, magnetic films, and all-new write and read heads.
IEEE's latest HDD development roadmap spans 2022 to 2037 and covers 15 years of hard drive evolution. The arrival of HAMR in 2024 will play a pivotal role in the increase in HDD capacity (even though Western Digital has managed to stay competitive with Seagate's HAMR HDDs using a set of its technologies) over the next few years. IEEE engineers expect HDDs to leapfrog to 40TB in 2025 and 60TB in 2028, doubling capacity from 30TB in 2024. By 2037, there will be 100TB of storage space, according to IEEE.
To get to those extreme capacities, HDD makers will have to increase the areal density of their platters steadily. To get to 40TB per drive, they will have to get to 2 TB/inch^2 in 2025 and then to over 4 TB/inch^2 in 2028 to build 60TB HDDs. By 2037, areal density will grow to over 10 Tb/inch^2. Increasing areal density will necessitate the use of new media, magnetic films, and all-new write and read heads.
Disc two (Score:2)
With Sci-hub claiming 77 TB in 2021 ( https://www.reddit.com/r/DataH... [reddit.com] ), by 2030 the disk capacity will match that of Futurama's Mars University, where non-fiction fits on a single disc ( https://theinfosphere.org/Mars... [theinfosphere.org] ).
Re: (Score:2)
The math doesn't seem to line up (Score:3)
40TB per drive, they will have to get to 2 TB/inch^2
implying 20sq in., but
4 TB/inch^2 in 2028 to build 60TB HDDs
implying 15 sq. in.
then
By 2037, areal density will grow to over 10 Tb/inch^2
implying 10 sq. in.
I'm confused.
Re: (Score:2)
Bandwidth better catch up (Score:4, Interesting)
I don't want to be backing up my drive for a week, we will either need to start seeing internal HD interfaces with ACTUAL write speeds approaching 1TB/sec or we'll have to buy two or more with a LOT of internal suspension to avoid data corruption and a special bay that we can just pull one out of and push another into, so we can just swap a new one in to replace the old one. And of course we'd need to deal with a second internal drive and tech inside with Raid 1, and the new drive would have to be stable long enough for the mirroring to finish.
Re: (Score:2)
Re:Bandwidth better catch up (Score:5, Interesting)
60TB at 100MB/sec is 7 Days, 15 Hours, and 15 Minutes, 7 Seconds which is terrifying (BTW, I did TB not TiB, same for MB).
I run much larger storage systems than this. I will order 1080 of these just to test them and place a big order if they prove reliable. We keep 3 copies of every file geolocated across 120 countries. With 108 drive wide stripes, we see pretty good performance.
Backup just isn't possible. We do have considerably more tape than disk, but that's just cold data. We have legal requirements to retain about 2TB of new data for 25 years and make it accessible at 100Gb/sec for warm data. Hot data is 12TB/sec.
If I work in enterprise or cloud, I suspect I'd be using these drives only for cold storage in 3x3 3xRAID6 by 3xphysical locations. So, 9 drives to provide 60TB of reliable write-once data sounds pretty good
Re: (Score:2)
This isn't for you.
60TB at 100MB/sec is 7 Days, 15 Hours, and 15 Minutes, 7 Seconds which is terrifying (BTW, I did TB not TiB, same for MB).
I run much larger storage systems than this. I will order 1080 of these just to test them and place a big order if they prove reliable. We keep 3 copies of every file geolocated across 120 countries. With 108 drive wide stripes, we see pretty good performance.
Backup just isn't possible. We do have considerably more tape than disk, but that's just cold data. We have legal requirements to retain about 2TB of new data for 25 years and make it accessible at 100Gb/sec for warm data. Hot data is 12TB/sec.
If I work in enterprise or cloud, I suspect I'd be using these drives only for cold storage in 3x3 3xRAID6 by 3xphysical locations. So, 9 drives to provide 60TB of reliable write-once data sounds pretty good
This is why you're not backing up an entire drive all at once. You back it up once, then back up changes and dedup.
We regularly deal with 60TB LUNs in the virtualisation world, 60TB is not even the biggest, it's not unusual to see a VSAN in the 100's of TB. I'm certain storage and backup admins could scoff at such numbers too.
However I'm sceptical of seeing a 60 TB drive in the market in 3 years, the largest one I can order from a store is 24 TB (apparently there's a 30TB Seagate out there). What I wa
Re: (Score:3)
I don't want to be backing up my drive for a week
You're backing up your drive for a week regardless. The data size is the data size, whether you move it on one drive or 6 is irrelevant. Sure you may say each individual drive is smaller so you only need to backup 10TB instead of 60TB , but to that I say with each individual size you multiply your chance of failure and will be sitting there moving those 10TB 6x more frequently.
If you want to backup less, then then only option is storing less data.
Re: (Score:2)
If you have multiple drives, you can read from and write to them at the same time. It is called RAID and is pretty common to set them up this way.
The time to read/write a single drive is determined by:
The rpm (rotations [or number of tracks] per minute) - this has not got any faster since I started following these things in the late 1900s.
The number of bits per track - this has increased a lot, and does increase data transfer speed.
The number of tracks per platter - this has increased by about the same as b
Re: (Score:2)
If you have multiple drives, you can read from and write to them at the same time. It is called RAID and is pretty common to set them up this way.
When the OP said "backup" I naturally assumed he'd be talking about rebuilding a RAID array. Point is data takes time to transfer. The only saving grace you have is less data.
The number of bits per track - this has increased a lot, and does increase data transfer speed.
This unfortunately has not been even an equal factor. With higher density came added complexity which forces some issue with transferring at the same speeds as the past. Look to something like Seagate's highest density drive that isn't SMR (since that wouldn't be a fair comparison) at 24TB it transfers at 280MB/s continuous. Look to th
Storage. (Score:2)
Most of us won't need to see this kind of size increase in storage devices. This is clearly for data centers and cloud computing, etc, etc..
I can't imagine squeezing more data into smaller spaces will be good for the life of a drive, so data backups will be even more important.
Re: (Score:2)
Most of us won't need to see this kind of size increase in storage devices. This is clearly for data centers and cloud computing, etc, etc..
I can't imagine squeezing more data into smaller spaces will be good for the life of a drive, so data backups will be even more important.
These extremely large HDDs are for cold storage, i.e., large amounts of mostly read-only, rarely accessed data. Like older items on social media sites that take a few seconds to appear when one happens to scroll down that far.
Any data that is frequently updated will need replication and rebuild, and that won't be practical with these large capacities.
Re:Storage. (Score:4, Informative)
> Most of us won't need to see this kind of size increase in storage devices.
Speak for yourself. My porn folder is almost full.
That's what they said about... (Score:1)
40MB drives back in 1990. Notice how that comment didn't age well?
Are HDDs still all made in one spot on Earth? (Score:2)
I guess these are for data centers and data hoarders.
Re:Are HDDs still all made in one spot on Earth? (Score:4, Informative)
>"Sorry I switched to Solid State and am so far glad for not having to listen to bearing whine or bad sectors suddenly popping up and growing"
Solid state drives are just wonderful. Small, low-power, ultra fast, reliable. They breathed new life into computers, for sure. I am a huge fan.
The problem is that if you have a lot of data to store, solid state is still way, way more expensive. So if you have the physical space and don't need incredible speed, spinning hard drives still have a major role. At least until SSD storage sizes increase a lot and the price goes down a whole lot.
A single 20TB hard drive, name brand, is now around $230. M.2 SSD is around $900 for just 8TB. So to meet that 20TB mark, you could do that in lower-priced 4TB M.2 drives, but you would need *5* of them at around $240 each, which is $1200, and something that can handle 5 M.2 slots. Now jump to 60TB and you would need just 3 20TB spinning hard drives, very doable. But 15 M.2 slots??
>"I guess these are for data centers and data hoarders."
Video. I guess you haven't tried to archive 4K video lately. It adds up really quickly. I am looking at storage requirements for security systems and if you have a good number of cameras and need to hold that for several weeks, SSD isn't even remotely reasonable.
Re:Are HDDs still all made in one spot on Earth? (Score:4, Interesting)
Solid state drives are just wonderful. Small, low-power, ultra fast, reliable. They breathed new life into computers, for sure. I am a huge fan.
Yes, but they have a nasty habit of losing data when not powered for extended periods of time (multiple years). So for backup purposes of data that may not be accessed for a long time, I would rather rely on magneto-optical hard magnetic drives. Maybe one day we'll see a special kind of SSD on the shelves that is built to keep its content for decades, even if not powered.
Re: (Score:1)
When you go past 8 TB, ignore M.2. Use enterprise-grade drives, which come in the U.2/U.3 factor, plus adapters. And you can already have 60 TB on a single QLC NVMe drive (Solidigm D5-P5336, available now for 7200 USD). Or, if you don't want QLC, there are multiple 30 TB SSD offerings from other manufacturers.
Re: (Score:3)
No bearing whine, but the bad sector was the whole NVME. It was only four and a half years old.
Backups are still needed, but you knew that.
Re: (Score:2)
Lots of purposes (Score:1)
What we nee
Re: (Score:2)
Right now I have Proxmox on my desktop computer with a Windows VM
Although I do like Proxmox, I wouldn't trust it for keeping data safe. After an update earlier this year, my own installation has gained a nasty tendency to blank out the disk labels of VM virtual drives at random during a reboot. Which means loss of the partition tables and non-booting VMs is a regular occurrence, and I have to keep around fdisk scripts for each VM to restore them when that happens.
INB4 "why not just use gpt?": Because an update to Proxmox around the start of the summer made booting a U
No thanks (Score:2)
Hard drives are annoying just ditch them so that manufacturers can focus own SSD.
Re: (Score:2)
Hard drives are annoying just ditch them so that manufacturers can focus own SSD.
They're annoying due to being slow. But flash drives are annoying due to being very expensive by comparison and in the case of nvme being pretty irritating to install in quantity.
I've got 28T of spinning storage and 5T of flash. That would be rather shockingly expensive in flash.
Predictions (Score:2)
How can they predict out to 2037? If they know it can be done, that means they know how. Why can't they release it now. I get it .. they need to build the manufacturing infrastructure. And that's fine, but that can explain things they are saying will be ready in say 3 or, max, 4 years. Beyond that it means they have no idea how to do it.
Re: (Score:2)
Or maybe I've seen people use extrapolation to assume something and it doesn't work. We should have fusion energy, room temperature superconductors, lunar bases, and stage IV cancer cures by now.
Will they be able 2 make em fit in 2.5" ext.cases? (Score:2)
Huge storage sized SSDs are still too expensive compared to HDDs. USB flash drives are too slow and unreliable. :(
Re: (Score:1)
Crazy progress! (Score:1)
Re: (Score:2)
1TB? I remember when we put a 120 MB Maxtor full-height MFM drive in a customer's Novell file server in 1987. The drive was around $3000 in 1987 dollars IIRC, and we just gazed upon it in awe, not being able to imagine even coming close to filling it up. Nowadays, I don't even want to be bothered with 128GB flash drives that hold more than a thousand times more than that and cost less than lunch. It's just ridiculous now.
Shingle BS (Score:2)