Samsung is about to make 4TB SSDs and mobile storage cheaper
A couple of years ago, Samsung launched its first 4TB solid state drives, which might as well not have existed given their $1,499 asking price. Today, the company announces the commencement of mass production of a more — though it's too early to know exactly how much more — affordable variant with its 4TB QLC SSDs. The knock on QLC NAND storage has traditionally been that it sacrifices speed for an increased density, however Samsung promises the same 540MBps read and 520MBps write speeds for its new SSDs as it offers on its existing SATA SSD drives.
Describing this new family of storage drives, which will also include 1TB and 2TB variants, as consumer class, Samsung will obviously aim to price them at a level where quibbles about performance will be overwhelmed by the sheer advantage of having terabytes of space. Any concerns about the reliability of these drives should also be allayed by the three-year warranty promised by Samsung. The launch of the first drives built around these new storage chips is slated for later this year.
What's the endurance of QLC NAND again?
Also at Engadget.
Related: Toshiba's 3D QLC NAND Could Reach 1000 P/E Cycles
Samsung Announces a 128 TB SSD With QLC NAND
Micron Launches First QLC NAND SSD
Western Digital Samples 96-Layer 3D QLC NAND with 1.33 Tb Per Die
Related Stories
While QLC NAND is predicted to have as low as 100 program/erase cycles (endurance), Toshiba has "targeted" 1000 cycles for its upcoming 3D QLC NAND products:
Toshiba last week announced its first 3D NAND flash memory chips featuring [the] QLC (quadruple level cell) BiCS architecture. The new components feature 64 layers and developers of SSDs and SSD [controllers] have already received samples of the devices, which Toshiba plans to use for various types of storage solutions.
[...] Besides [its] intention to produce 768 Gb 3D QLC NAND flash for the aforementioned devices, the most interesting part of Toshiba's announcement is [the] endurance specification for the upcoming components. According to the company, its 3D QLC NAND is targeted for ~1000 program/erase cycles, which is close to TLC NAND flash. This is considerably higher than the amount of P/E cycles (100 – 150) expected for QLC by the industry over the years. At first thought, it comes across [as] a typo - didn't they mean 100?. But the email we received was quite clear:
- What's the number of P/E cycles supported by Toshiba's QLC NAND?
- QLC P/E is targeted for 1K cycles.
Endurance miracle putting QLC on par with TLC, or idle talk about a product that won't be out for 1-2 years?
[Ed. note: If you're wondering what QLC NAND is, here's a quick primer.]
Additional Coverage: The guru of 3D
Samsung will use QLC NAND to create a 128 TB SSD:
For now, let's talk about the goods we'll see over the next year. The biggest news to come out of the new Samsung campus is QLC flash. Samsung's customers set performance and endurance specifications and don't care about the underlying technology as long as those needs are met. Samsung says it can achieve its targets with its first generation QLC (4-bits per cell) V-NAND technology.
The first product pre-announcement (it doesn't have a product number yet) is a 128TB SAS SSD using QLC technology with a 1TB die size. The company plans to go beyond 16 die per package using chip stacking technology that will yield 32 die per package, a flash industry record.
NAND revenue has increased 55% in one year.
Previously: Seagate Demonstrates a 60 TB 3.5" SSD
Toshiba Envisions a 100 TB QLC SSD in the "Near Future"
Western Digital Announces 96-Layer 3D NAND, Including Both TLC and QLC
Toshiba's 3D QLC NAND Could Reach 1000 P/E Cycles
Intel and Micron's jointly-developed 3D QLC (4 bits per cell) NAND memory is featured in a new Micron enterprise SSD, the 5210 ION. The drive will have a capacity of 1.92 TB, 3.84 TB, or 7.68 TB, and a write endurance of less than 1 drive write per day (possibly as low as 0.1 DWPD):
The cost reduction brought by QLC NAND is a much-awaited advance for enterprise storage. Most NAND flash manufacturers have started sampling QLC NAND within the past year, generally built on the same 64-layer 3D NAND processes that current-generation TLC NAND uses. Micron has previously shown wafers of 512Gb 64-layer QLC when announcing the addition of QLC to their roadmap, but today they are also announcing a 1Tb 64L QLC part—the first 1Tb memory chip to hit commercial availability. That 1Tb part is organized as four planes that can be processing I/O commands in parallel, compared to two planes for previous Intel/Micron NAND parts. This helps offset most of the performance loss associated with increasing per-die capacity. Thanks to the "CMOS under the array" design of Intel/Micron 3D NAND, the extra peripheral circuitry requried by doubling the number of planes doesn't add much to the overall die size.
It was initially feared that QLC write endurance would be low enough that drives would need to be treated more or less as write-once, read-many (WORM) devices, requiring careful handling on the software side. With multiple manufacturers now rating their QLC NAND for around 1k P/E cycles, it is clear that QLC-based SSDs aren't too fragile and can handle many existing workloads without needing major software changes to reduce writes.
Micron is primarily marketing the 5210 ION SSDs as replacement for hard drives, rather than replacements for any existing tier of enterprise SSD products. In this role, the 5210 ION will have clear advantages in density (with 2-8TB per 2.5" drive) and performance. QLC NAND only provides incremental improvements to cost, so the 5210 ION won't be matching 7200RPM hard drives for price per GB, but 10k RPM drives will probably be feeling the pressure, especially from TCO calculations that take into account the power efficiency advantages of SSDs.
The next generation of Intel/Micron 3D NAND will have 96 layers, potentially using string-stacking to combine two 48-layer dies. After that, Intel and Micron will go their separate ways.
Western Digital Begins to Sample QLC BiCS4: 1.33 Tbit 96-Layer 3D NAND
Western Digital has started sampling its 96-layer 3D NAND chips featuring QLC architecture that stores four bits per cell. The chip happens to be the world's highest-capacity 3D NAND device. The company expects to commence volume shipments of this memory chip already this calendar year.
Western Digital's 96-layer BICS4 3D QLC NAND chip can store up to 1.33 Tb of raw data, or around 166 GB. The IC will be initially used for consumer products Western Digital sells under the SanDisk brand, so think of memory cards (e.g., high-capacity SD and microSD products), USB drives, and some other devices. The manufacturer expects its 3D QLD[sic] NAND memory to be used in a variety of applications, including retail, mobile, embedded, client, and enterprise, but does not elaborate on timing at this point.
The 1.33-Tb BICS4 IC is Western Digital's second-gen 3D QLC NAND device. Last year the company announced its BICS3 64-layer 3D QLC chips featuring a 768 Gb capacity, but it is unclear whether they have ever been used for commercial products. Meanwhile, it is clear that the device was used to learn about 3D QLC behavior in general (i.e., endurance, read errors, retention, etc.)
[...] What is noteworthy is that officially the BiCS4 range was to include both TLC and QLC ICs with capacities ranging from 256 Gb to 1 Tb, so the 1.33 Tb IC is a surprising addition to the lineup which signals Western Digital's confidence of its technology.
Recent products have been using 512 Gb per die NAND, with 768 Gb and 1 Tb on the horizon. Samsung's announced 128 TB SSD was supposed to use 1 Tb 3D QLC dies, so ~1.33 Tb dies could bring that capacity to about 170 TB. Given a couple more generations of NAND or some fancy die/package stacking, and we will probably see a 1 petabyte SSD.
Samsung Shares SSD Roadmap for QLC NAND And 96-layer 3D NAND
At Samsung's Tech Day event today in San Jose, the company shared their SSD roadmap for transitioning to 96-layer 3D NAND and introducing four bit per cell (QLC) NAND flash memory. Successors have been named for most of their current SSDs that use three bit per cell (TLC) NAND flash and are being updated with 96-layer 3D TLC, and new product lines using QLC NAND have been introduced.
[...] The enterprise SAS product line is not seeing any major changes to performance or available capacities, but the update from the PM1643 to the PM1643a does improve random write performance by about 20%. The largest model remains 30.72TB. The high-end enterprise NVMe drives are getting a major controller update that brings PCIe 4.0 support in addition to the NAND upgrade. This allows for much higher performance across the board, most notably with sequential read speeds reaching 8GB/s on the new PM1733 compared to 3.5GB/s on the PM1723b. The maximum available capacity has caught up to the SAS product line with the introduction of a 30.72TB model.
[...] Samsung also mentioned that in Q2 2019 they are planning to introduce a higher-performing 512Gb QLC die to complement their current 1Tb die. Samsung compared the performance of this new 512Gb die against an unspecified competitor's 1Tb QLC, claiming that Samsung's high-performance QLC will have 37% lower read latency and 45% lower program latency.
[...] The first products featuring the second generation of Samsung's low-latency Z-NAND flash memory will be the SZ1733 and SZ1735, high-end enterprise NVMe SSDs that differ primarily in the amount of overprovisioning. Samsung has announced that their second generation of Z-NAND will include a MLC version, but these drives are using the SLC version. Like the TLC-based PM1733, the new Z-NAND SSDs will also feature dual-port capability and PCIe 4.0 support. Sequential reads of up to 12GB/s are claimed, but this product line is all about random I/O, which Samsung hasn't detailed yet. Samsung demoed a 4TB model, significantly larger than the 800GB maximum for the first-generation SZ985.
Z-NAND (PDF) has lower latency than normal NAND, and could be compared to Intel and Micron's 3D XPoint.
Related: Western Digital and Samsung at the Flash Memory Summit
Samsung Announces a 128 TB SSD With QLC NAND
Samsung Announces Production of 1-4 TB Consumer 3D QLC NAND SSDs
(Score: 2) by requerdanos on Wednesday August 08 2018, @01:33AM (8 children)
If Samsung could warranty the customer's data, that would no doubt be true. Since Samsung is only going to warranty the drive, however, the warranty isn't going to be allaying many reliability fears, especially those concerned with integrity of data.
(Score: 1) by khallow on Wednesday August 08 2018, @04:36AM
What it does mean is that they'll lose money very rapidly if their drives have a high failure rate over the warrantee period. At a 5% failure rate per month, for example, two years in they would expect to replace more drives than they originally sold.
(Score: -1, Troll) by Anonymous Coward on Wednesday August 08 2018, @05:22AM (6 children)
So, if a company warranties data, and the device fails, and then what? How is the firm supposed to compensate the buyer?
That's right. You are a moron.
(Score: 2) by requerdanos on Wednesday August 08 2018, @02:21PM (4 children)
I used to sell surge protecting power strips guaranteed to stop lightning to my computer repair clients.
Now, of course, no box full of wires is going to stop lightning. If a lightning-induced surge enters at one end, it's going to come out the other end, regardless of the arrangements of the "surge protecting" components within.
How did the firm compensate the buyers, then, right? By compensating for lightning-related losses up to a million dollars. So, no, the data did not magically reappear (sorry for raising your hopes), but there was (theoretically) a way to make up the value of the lost data.
That doesn't appear to be the case with these large SSDs (looks like you would just get a replacement, empty SSD). So I pointed out that it doesn't work that way.
The liar assumes that everyone else will lie; the thief assumes that everyone else steals. I guess the moron assumes that no one's smarter.
(Score: 0) by Anonymous Coward on Wednesday August 08 2018, @03:42PM (3 children)
You are simply raising the liability limit from the cost of the device to some arbitrary numbers. Basically you are arguing that the device should comes with an insurance. That simply raises the price for everyone when only a small minority want it and is willing to pay for it.
That "moron" comment was uncalled for, and that's my bad, but yours is a dumb idea.
(Score: 3, Informative) by requerdanos on Wednesday August 08 2018, @04:09PM (2 children)
I appreciate your participation, but I don't think you are following the meaning of the thread.
Because the devices do not come with data insurance, a warranty on the device is no reassurance.
Both the author(s) of TFA and you yourself seem to have trouble grasping this.
The fix for this is not to provide insurance for the device--your interpretation, not mine, and agreed, a dumb one--but to recognize that because the devices do not come with something like data insurance (why would they?), a warranty on the device is no reassurance.
This is because a warranty on the device itself is no reassurance if you value your data but question the reliability of the device upon which you decide to store it. In contrast to what TFA/TFS says, which is the opposite. And which is wrong.
(Score: 0) by Anonymous Coward on Wednesday August 08 2018, @04:47PM
You are right; I misread your original comment. I chased up the wrong tree.
(Score: 2) by FatPhil on Thursday August 09 2018, @03:21AM
Somewhat bizarrely, it's a hedge against a counter-quality bet that they're making.
They're implicitly betting that the drive fails and you lose your data. When it fails, you lose, as you were unwittingly gambling pro-quality. However they're explicitly betting against themselves with the smaller bet that permits you to win in the failure case.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 2) by FatPhil on Thursday August 09 2018, @03:11AM
B: X is impossible
How the fuck is that a counter-argument?
Funnily, X is possible, so you're both logically and factually wrong. System/solution vendors can sell a "system" or a "solution", rather than just a small piece of hardware. Sure, you pay loads for them to manage it, but you get the guarantee you demanded.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 0) by Anonymous Coward on Wednesday August 08 2018, @04:33AM (5 children)
Not relevant maybe with amazon, but HDD sizes, selection and price have stalled where i live.
the "growth" curve of HDD has been very flat the last few years whilst the SSD only has to fight
"being appropriated as a m$ device, like win-printers, win-scanners etc.(*)" but overall have had a step
"growth curve".
maybe if i hold out another two years ... the next YUGE replacement HDD will in fact rather be a YUGE SSD?
(*) only works if windows is installed and has drivers (that aren't obsoleted yet).
(Score: 4, Insightful) by takyon on Wednesday August 08 2018, @06:58AM (4 children)
Just looking around I see a 10 TB Seagate IronWolf HDD for $280, and 8 terabyte internals/externals ranging from $150-200. Even if you doubled or tripled those prices due to your country, it's going to be a lot cheaper than the equivalent in SSDs. For some years.
The new higher capacities (such as 12 and 14 TB right now) have debuted in enterprise products first, however Seagate is promising a consumer 14 TB drive this year [anandtech.com]. It might debut around $600-$800 (I may be guessing too high) before declining to reach in the ballpark of $0.02/GB. The Samsung 850 EVO 4TB SSD is around $1,100 right now, still around 10x the price per GB of on-sale HDDs. Maybe the new QLC SSDs will cut that in half.
At $0.10 to $0.15/GB, a 1 TB SSD would be $100-$150. This is where hard drives start to look bad because the smallest hard drives don't get much cheaper than $40. And the SSD is a product that greatly increases the performance of the computer compared to an HDD. Most people could live with 1 TB as their primary drive.
You hold out for 2 years, and you'll see more use of QLC, and more layers (96, maybe 128). You should see another 30% decline in $/GB. QLC should be scrutinized for endurance/longevity though.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by requerdanos on Wednesday August 08 2018, @03:58PM (3 children)
Since the beginning of time as it's counted in the home computer and PC market until the foreseeable future, the ability to store data on fast chips or in magnetic bubbles or some such instead of on hard drives has been more expensive, in fact.
CPUs get faster, memory gets faster, communications busses get faster, input/output interfaces get faster with time. Over the decades that we have had mini- and personal computers, these things have gotten faster and faster at a rate ranging from impressive to terrific.
Hard drives, on the other hand, are still very slow (and high-latency) relative to all the other data-moving and holding components and technologies. They get faster with time, sure, but they are still the device that has the greatest potential to create a performance bottleneck by orders of magnitude of slowerness.
It used to be that the single upgrade component that would make the most performance difference in a computer was adding memory--which would reduce reliance on a swap partition/file on the slow mechanical storage device. But now, it's adding an SSD. Boot is faster, programs start much faster, the storage is suddenly fast like everything else is. There are still bottlenecks, but they're much smaller and more complex--this is the last bit of low-hanging fruit.
The original computers that were built worked from moving parts, not silicon, and hard drives are the last holdover from that slow, clunky, but technologically exciting time. SSDs are not as fast as DDR-n memory, for example, but they represent a complete breakaway from the mechanical and into the solid-state.
So I guess it's fair that they cost a little more.
(Score: 3, Interesting) by takyon on Wednesday August 08 2018, @04:43PM (2 children)
This point is already covered in my comment, although you expanded on it.
---
There are different views on whether this will stay the case.
Updated IDC SSD forecast sees 44 per cent capacity CAGR 2016-2021 [theregister.co.uk]
If price/capacity was only 2x higher than HDDs rather than an order of magnitude, then HDDs would become a niche product fast. Especially given that SSDs are going to reach much higher capacities than HDDs can in the near term. Producing a 100 TB HDD would be incredibly difficult within the next 5 years, but producing a 1,000 TB SSD would not be that hard.
Hard Drive Cost Per Gigabyte [backblaze.com]
(Also see graph [backblaze.com]) So Backblaze confirms what OP was saying. There has been a flattening, which can be traced right back to the Thailand floods.
However, when WD announced MAMR [anandtech.com] (Seagate later confirmed it is sticking with HAMR), WD projected [anandtech.com] that HDDs could maintain an order of magnitude price/capacity advantage over SSDs (e.g. $10/TB for HDDs vs. $100/TB for SSDs).
The problem is that their projection graph looks very unrealstic. They forecast $/TB for SSDs declining 16-20% over 5 year periods. But it's probably more like 30% over 2 year periods from what I've read elsewhere. They have QLC hanging out in 2022-2028 when it will actually be shipping next year. I doubt the future of technologies over 4-bits-per-cell (although there has been interest [theregister.co.uk]) but by the time 2022 rolls around the amount of layers could be drastically increased to past 128 [soylentnews.org]. There has also been a lot of innovation in stacking many dies/packages [soylentnews.org]. Layers of layers. Which is why a 1 petabyte SSD is very feasible in the near future. Although massively increasing capacities doesn't necessarily result in a massive decrease in $/TB, it will help.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Wednesday August 08 2018, @07:58PM (1 child)
thanks for all comments.
the jab at the country was just to point out that some countries manufacture products andthen just export them.
it is super strange that the HDD factories here in thailand don't have a small table outside the factory where
each and every HDD they manufacture can be bought ...
it would be strange also, if, living in 'murika, you would have to order that "special" AMD or intel cpu from alibaba, non?
tho i guess it's hard to make a case for that table-outside-the-factory if it is surrounded by low lying rice marshes ;
owning a HDD doesn't magically make your rice grow faster ...
anyways, i must admit that the WD HDDs i have lying around are still working as promised, so thumbs-up for reliability.
on the other hand i have had some early SSDs in the 20-40 GB range die on me.
also one last point: energy usage.
the 2.5" HDD are about in the same ball park as their 2.5" SSD brethren; not so the 3.5" big brother.
if you have a few (5?) 8TB 3.5" HDD then that combo is using as much as one of those big CPU in the >80 W TDP range?
if a 500GB SSD uses as much energy as a 50TB SSD ... then ...
(Score: 2) by takyon on Wednesday August 08 2018, @08:54PM
I think the big SSDs tend to use more energy than their smaller counterparts (due to more packages... compare to the M.2 [wikipedia.org] chewing gum-like form factors). However, when it comes down to a home user with an SSD, even a power-hungry SSD will probably use less energy than an HDD since it completes random operations very fast and is also faster at sequential access. So it ends up completing reads and transfers and going back to idling.
One recent advance has helped HDDs reduce power consumption somewhat: helium-filled drives. Which are apparently available to consumers [anandtech.com] these days.
Maybe multi-actuator technology [soylentnews.org] could also reduce power consumption in some scenarios.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]