Samsung Starts Mass Producing an SSD With Monstrous 30.72TB Capacity (betanews.com)

Brian Fagioli, writing for BetaNews: Samsung says it is mass producing a solid state drive with monstrous capacity. The "PM1643," as it is called, offers an insane 30.72TB of storage space! This is achieved by using 32 x 1TB NAND flash. "Samsung reached the new capacity and performance enhancements through several technology progressions in the design of its controller, DRAM packaging and associated software. Included in these advancements is a highly efficient controller architecture that integrates nine controllers from the previous high-capacity SSD lineup into a single package, enabling a greater amount of space within the SSD to be used for storage. The PM1643 drive also applies Through Silicon Via (TSV) technology to interconnect 8Gb DDR4 chips, creating 10 4GB TSV DRAM packages, totaling 40GB of DRAM. This marks the first time that TSV-applied DRAM has been used in an SSD," says Samsung.

Comments

Price?

[Anonymous Coward]

And how much will this thing cost?

Yeah, that's what I thought.

Re:Price?

[Sigma 7]

And how much will this thing cost?

It's hard to find prices for these devices, almost as if there's some secondary market where prices are negotiated without being displayed publicly. But from what I can tell, it's seems to be sent towards enterprise-tier, thus you'd have to be a big company just to get one.

If existing prices are a good way to calculate price of the 30TB whopper, it's likely to cost $14250 each (using $/TB from a Samsung MZ-76P4T0BW). Currently cheaper to get a three pack of spinning 12TB drives (totalling $1350).

Re:Price?

[EvilSS]

Currently cheaper to get a three pack of spinning 12TB drives (totalling $1350).

No one is really buying SSD in the enterprise based on it's price vs HDD.

Re:Price?

[dgatwood]

No, but a lot of folks in enterprise will not by it based on its price compared with spinning rust. You can afford to spend a lot of time structuring your databases and similar so that frequently accessed data stays warm in a cache and still come out cheaper than the price of moving all your company's data to SSDs.

Re:Price?

[DickBreath]

> And how much will this thing cost?

Wrong question.

How many simultaneous pr0n streams can it sustain?

Re:Price?

[ShanghaiBill]

You can afford to spend a lot of time structuring your databases ...

The extrapolated cost of this 30TB drive is less than the burdened monthly cost of a single engineer. Furthermore, if disk I/O is your bottleneck, then moving your DB from HDD to SDD is likely to give you far more speedup than fiddling with your schema.

Re:Price?

[Nostalgia4Infinity]

In my experience running IT for a large department the cost saved in not having to replace drives, and worse, restore data is worth the cost of SSDs, probably just in labor costs but also the additional peace of mind.

Also:
"For most corporate workloads, the acquisition cost of flash storage is still significantly greater than for HDD storage. However, when operating costs are factored in, the TCO for SSDs may actually already be lower than for equivalent HDD arrays. Use of SSDs reduces data center costs for power, cooling, floor space, rack space, and maintenance. And as SSD purchase prices continue to fall, the TCO disparity can only grow greater over time."

https://www.zadarastorage.com/...

Re:Price?

[dgatwood]

The extrapolated cost of this 30TB drive is less than the burdened monthly cost of a single engineer.

And if your company is small enough that a single 30 TB SSD is enough to meet your entire storage needs, then you probably aren't big enough to be so desperate for speed that you would buy a 30 TB SSD. I'm expecting most of these will be used for the most frequently accessed data for companies that are Google-scale, not companies whose total data capacity needs are rivaled only by my home RAID array. :-)

Furthermore, if disk I/O is your bottleneck, then moving your DB from HDD to SDD is likely to give you far more speedup than fiddling with your schema.

But "fiddling with your schema" so that tables that get hit frequently are indexed on SSDs or DRAM drives (and, if necessary, stored in their entirety on those drives) means you get most of the speed win without the expense of storing your *entire* data collection on an SSD. The more you can separate out the frequently accessed data from the rarely accessed data, the easier it is to keep your costs sane. And the bigger your total storage needs, the more that design decision matters.

Re:Price?

[fahrbot-bot]

Currently cheaper to get a three pack of spinning 12TB drives (totalling $1350).

No one is really buying SSD in the enterprise based on it's price vs HDD.

Sure, but it's a factor for the non Galaxy class Federation Starships.

Re: Price?

[Zero__Kelvin]

He expicitly stated that nobody in Enterprise buys SSDs. It is hard to imagine why you would decide he didn't mean that. If he wants to change his statement he can, but you can't change it for him.

Re:Price?

[JMJimmy]

Bring it down to $300 and I'll buy one.

I mean SSDs been available for ~10-26 years now, depending on how you define "available", you'd think they'd have figured out how to get the $/GB below that of HDDs long ago but it's maybe 6-12 months away before hitting that milestone, let alone advancing to the $/TB milestone.

Re:Price?

[Bitmanhome]

It's got 40GB of RAM, so pretty much all of them.

Re: Price?

[BronsCon]

No, he stated that the price difference isn't a factor. The meaning you claim his statement held would have been worded "No one is really buying SSD in the enterprise, based on it's price vs HDD." Note how that's two separate clauses; that comma does change the meaning to what you state, but that comma wasn't present in the original comment.

Re:Price?

[angel'o'sphere]

A day of a competent DBA costs about $1000.
You want to tell me a competent DBA will outperform on an HDD a less dompetent one on an SSD?
Dream on ...

Re: Price?

[kenh]

You can't have them in you workstation, they make too much noise and are too slow for gaming.

WTF are you smoking?

How "loud" are 3 1/2" drives?

SATA III drives are "too slow" for gaming?

I watch truckloads of SATA HDD roll out of my local Fry's and Microcenter, I don't think everyone buying HDD has a server farm at home.

Re:Price?

[dgatwood]

No, what I mean is that a good DBA can make the difference between needing to store everything on an SSD and needing only to store indices and frequently accessed data, and the rest of the performance problems can be handled with sharding for a lot less than the cost of storing everything on an SSD.

Re:Price?

[EvilSS]

Like I said: No one is really buying SSD in the enterprise based on it's price vs HDD. You won't save Capex buying the same size array going SSD vs HDD. However, the price for SSD storage is at the point that the advantages (performance, density, and reliability*) outweigh the price difference. You can get enterprise SSD solutions for under $0.30/gb from multiple vendors. At some point it's cheaper to go SSD than to invest in "spending a lot of time" trying to make it faster on HDD. I'd say probably close to 70% of the storage systems my company has sold in the past 12 months are SSD, closer to 90% in the past 6 months.

*I've spoken with our support guys, customers, and with field techs (not SE's, but guys who actually go out on service calls) from several vendors and they are all saying about the same thing: drive failure rates are much lower than those with spinning disks on the new SSD systems. I'm sure someone will pipe up with "But SSD's wear out!!! BLARG!!" That's just not what we are seeing when we looked at the SSD wear rates vs traditional drive failures in production environments.

Re:Price?

[EvilSS]

What else are you going to use them for? You can't have mechanical disks in laptops or other portable devices, they just break. You can't have them in you workstation, they make too much noise and are too slow for gaming. The only remaining application for mechanical drives is in servers.

Well you can use them portables they just suck because you usually end up with 5400rpm drives and no one wants them. They work fine in workstations and many have them as secondary drives, booting and loading frequently used apps off of a smaller SSD because large SSDs are still somewhat expensive (for now, this is rapidly changing) . Also most games don't really benefit from faster storage. Some do, but most don't.

As for servers, nope, that's going away too. I just ordered up a bunch of Cisco UCS blades for a customer and guess what they boot from? No, not SSDs. SD Cards. Yep, a pair of SD Cards in RAID 1. They stick into the side of the blade. When most physical servers going in these days are just metal for hypervisors or compute nodes for some clustered task, putting drives in them isn't all that important. You just need them to boot reliably. So most of the storage is happening on SAN and NAS and those are going pure SSD at an ever increasing rate as SSD costs drop year over year.

I'm afraid in another decade or so the spinning disk is going to be a niche legacy product relegated to old systems and oddball situations where they might make more sense to use. One day you will be able to pull one out of a dusty box in the attic and show your grand kids how primitive data storage used to be.

Re:Price?

[EvilSS]

At this point it's really demand that's keeping the cost up. Between NAND and RAM, the fabs are all running full tilt and the memory cartel of Samsung, Micron, and Hynix doesn't seem all that interested in opening up new capacity. It's still a ways to go before it can hope to match HDD in price per GB, but it's not purely a technical hurdle anymore. I think you will see HDD demand start dropping in the coming year, and eventually that will push HDD prices up. I suspect in another decade or so SSD and HDD prices will reach parity, or at least get close enough that it won't matter.

Re:Price?

[EvilSS]

ugh, that should read " I think you will see HDD demand start dropping in the coming years"

Re:Price?

[Kjella]

And if your company is small enough that a single 30 TB SSD is enough to meet your entire storage needs, then you probably aren't big enough to be so desperate for speed that you would buy a 30 TB SSD. I'm expecting most of these will be used for the most frequently accessed data for companies that are Google-scale, not companies whose total data capacity needs are rivaled only by my home RAID array. :-)

I suppose it's what you do. To a video production company 30TB is nothing. For random document storage it's quite a bit. For a database/source code repository it's pretty huge. We have quite a bit of database data that we need to keep around for archive purposes that are almost never used again, but in the end the relative savings of HDD vs SSD in a relatively expensive storage unit that still needs backups etc. and the resources to partitioning and archive things it didn't add up so we're all SSD now. It's a bit like starting to go through your photo collection to see if you can delete 10MB jpgs when you can go down to the store, buy a 8TB archive disk and store 800000 more. If you're doing it simply to save space and not for house cleaning it's not worth it.

Re: Price?

[Dog-Cow]

You are a clear liar. Why don't you go find a nice deep pit to fall into?

Re:Price?

[TheRaven64]

I bought my first flash SSD in 1994 (I think, maybe 1995). It was for my Psion Series 3, had a capacity of 128KB, and cost £30. It was a single cell, so you could write to it but not erase without erasing the entire disk, so you needed to do a backup and restore to replace all of the data. At about that time, a 1GB hard disk cost around £100-150 (I don't remember exactly what, but I remember being amazed when the price dropped to under 10p/MB a little bit later). Let's say £120, for the sake of argument. That meant that the cost of a hard disk, per GB, was about 8,000 times less than the cost of flash.

It's been a couple of years since I last bought any disks (my last two laptops have had SSDs, but the newer one is now over four years old and has a 1TB SSD). It looks as if 500GB SSDs now cost about £120, which is about the same price as a 5TB hard disk. That's a difference of about a factor of 10 in price - a big jump from a factor of 8,000. I'd expect that gap to close pretty soon.

Perhaps more important is the bottom end. You can get a 120GB SSD for about £30, or a 500GB hard disk for the same price. Note that now we're down to a factor of four difference in price. For a lot of users (e.g. corporate desktops where important files are stored on a server), the difference in utility between 120GB and 500GB capacity isn't that great, but the difference in performance between SSD and hard disk is a lot more noticeable, so it makes sense to go with the SSD.

This is the real problem for hard disk: if enterprise, mobile, and low-end desktops move to SSDs, where do you get the volume to pay for R&D on the next generation?

Re:Price?

[JMJimmy]

If you graph the history of HDDs over the years there haven't been any significant advancements since 2011. We're basically at the 10nm theoretical limit right now (12TB). ODS showed promise for Petabyte storage but hasn't materialized commercially, despite being around since 2009. At this point it seems like profiteering is taking over from rapid advancement.

Re: Price?

[Zero__Kelvin]

Nice of you to try and get your mom the D, but I ain't touching that stank hoe.

Re:Price?

[EvilSS]

The do wear out, but only on writes, and the controllers are smart enough to balance it out well and reserve spare capacity for cells the fail early. And in a lot of real world workloads writes as a percentage or stored data is fairly low.

Yea not saying that they won't wear out, but there seems to be this recurring theme here and on other tech forums that SSDs are unreliable compared to HDD because of their limited write life, but it's just not what we are seeing in real world use.

Before you start drooling

[the_skywise]

Because it is intended for the enterprise and uses the Serial Attached SCSI interface, it is unlikely that it will be sold in any consumer retail channels.

But soon...

Re:Before you start drooling

[sexconker]

SAS/SCSI needs to go away.

I'm sick of "enterprise" gear being a shitty HBA to a shared storage device that slows you down (both speed and latency from the HBA, and the latency and shared aspect part of the storage device).

The world has spoken - NVMe PCIe flash isn't just the way of the future, it's the way of today.
But for servers I want it in a hot swap drive bay, and I need it to be redundant. There are only a handful of RAID controllers out there that will support such a thing. (And as far as I know, you're limited to 8 lanes of PCIe 3.0 with the latest "tri-mode" shit.)

Can we please get more support for NVMe PCIe storage in RAID controllers and backplanes? I've been actively waiting for this shit for about 2 years, and it's still a rarity and a mess. SAS/SCSI is dead. Let's bury it already. Fucking Epyc-based systems have 128 lanes of PCIe 3.0! Why don't we fucking use them?

Give me a backplane with 8 hot swap bays for 2.5" drives and pipe that shit over a 32 lane PCIe link (yes, it goes to 32, not just 16) to a fast RAID SOC. Each bay can get 4 lanes. And if you really want to be smart, you'll fucking let people connect M.2 devices on the backplane, not just U.2 devices. Because there are like 7 total U.2 devices in the world and they're all the same fucking thing as the M.2 equivalent in a 2.5" case but with a much larger price. Fuck that. I'll run servers on 960 Evos (or 960 Pros if I'm feeling paranoid). A M.2 to U.2 adapter, a 2.5" enclosure, and some thermal pads are all you need to get an affordable, fast, and reliable U.2 drive.

Re:Before you start drooling

[swb]

I'm not sure the HBA is really the problem. I mean, you want I/O offload to some other CPU for RAID and logical disk management and some kind of physical disk channel multiplexing.

There's no reason these things are inherently slow unless you're buying low end stuff. Obviously you'll get better raw I/O with direct NVMe off the PCIe bus to the CPU, but as soon as you get into RAID it gets more complicated. Either you use software RAID at the OS level or you have to add a RAID controller to your PCIe bus. The latter is actually worse than a HBA IMHO because the controller and the drives sharing the common PCIe bus slows the whole bus down for traffic that only the RAID controller needs to see.

With a HBA, you just write blocks to one device on your bus and you let the HBA write those to disk on a separate bus. There's no reason an HBA couldn't have its own PCIe bus for NVMe drives.

The reason I'm skeptical we'll see anything other than hot-swap NVMe on the main system bus is that storage seems to be lurching towards hyperconverged where you wouldn't have more than about 12 drives on a physical system anyway and it's all software RAID.

I think there is an evaporating market for device-dense storage devices.

Re:Before you start drooling

[sexconker]

The HBA itself isn't a problem. No individual piece of such a setup is. But the cost of the HBA and cables, dealing with cable routing in the rack, handling risers and chassis slot compatibility, etc. only to then have to deal with an external box (perhaps 2 for redundancy) that gives you less performance than you could get with direct attached storage is overall shitty.

If an HBA has its own PCIe bus, so can a RAID controller. There's no fundamental advantage there. The thing about software RAID is that ESXi doesn't support it.

And RAID doesn't just make sense for dense storage. It actually makes a lot more sense on lower end stuff. Consider someone who just wants 2 or 4 drives in RAID 1 for the fault tolerance or RAID 10 for the tolerance and performance/capacity. Telling them to run over an HBA to external box filled with drives represents a huge increase in cost and complexity. And if the RAID controller dies, you can have another in a matter of hours, or swap with a known good one. External storage devices take a lot longer to buy, receive, and physically install, let alone config. And it's a lot more expensive to have a second on hand just in case. If someone does need a lot of storage, then the jump to a dedicated storage box, and potential vendor contracts to support/maintain it, isn't as big.

Re:Before you start drooling

[swb]

I'm not arguing against RAID, just that if you really want it RAID HBAs have a lot of appeal.

Disk densities and shared storage have kind of nuked the external storage shelf market, which I agree was kind nuisance prone with its many cables and connectors.

The one good thing external cabinets had going for them was some portability between attached hosts. I have had situations where a host died and I had the ability to move the storage to another host's HBA RAID controller and pretty easily get the data active again.

It would be nice to see ESXi support software RAID. The fact they have embedded so much of vSAN into the kernel says to me they probably could do software RAID.

Re:Before you start drooling

[sexconker]

Yeah, I've been looking at vSAN, but it costs extra (of course, it's VMware, everything costs extra) and has some odd requirements.

AMD and Intel both have on-CPU RAID, and I'd love for Epyc and ESXi to get on this shit. With SSDs and redundant PSUs fed by redundant PDUs backed by UPS and generator power, I don't think I care about a RAID card's battery/supercap. Just let me use some fucking NVMe SSDs in a redundant, hot-swappable fashion. We already have all the pieces we need, but for some reason vendors don't want to put them together in any way that lets me stay on the official hardware compatibility list.

That 0.02 TB made the difference.

[140Mandak262Jamuna]

Originally I thought the capacity was 30.7 TB. I was like, meh! not impressed.

Then I looked at the fourth significant digit. it is 2. Yes, it is actually 30.72 TB. That 651 parts per million more than my what I originally thought. Now I am all ears, looking at it carefully, camping outside Alibaba container terminal to be the first one on the block to get it.

Very well done Dear Headline Writer, always provide very precise information. Next time, why stop with the fourth significant digit? You could be even more amazingly accurate and provide six, seven... why not go all the way to 11 significant digits! Most people have just 10 digits, so go for 11, that is a good number hard to beat.

Re:That 0.02 TB made the difference.

[Jason Levine]

They claim it's 30.72GB but all you really get is 30.1415926535GB of space.

Re:That 0.02 TB made the difference.

[Anonymous Coward]

That's 20GB you are making fun of. We used to dream of drives with a fraction of that capacity. (Uphill both ways, etc.)

Re:That 0.02 TB made the difference.

[JD-1027]

That 0.02 TB is 200 times the capacity of the first external hard drive I bought. I think I remember my first drive at 100 MB.

Re:That 0.02 TB made the difference.

[sconeu]

Meh. My first hard drive was 40MB -- special order, the machine's stock drive was 20MB.

So it's 500 times my first hard drive.

Re:That 0.02 TB made the difference.

[dgatwood]

640 kilobytes, but to be fair, that ought to be enough for anybody.

Re:That 0.02 TB made the difference.

[DontBeAMoran]

The "PM1643," as it is called, offers an insane 30.72TB of storage space! This is achieved by using 32 x 1TB NAND flash.

32 x 1TB NAND flash = 32TB or 30.72TiB.

Re:That 0.02 TB made the difference.

[HornWumpus]

Luxury...

Re:That 0.02 TB made the difference.

[AmiMoJo]

2^40 or 10^12 size terabytes?

Re:That 0.02 TB made the difference.

[DontBeAMoran]

I did not calculate that 30.72 number, I took it out of the headline thinking it was supposed to say 30.72 TiB instead of 30.72 TB.

Re: That 0.02 TB made the difference.

[Zero__Kelvin]

You have complained that accurate information was provided. Turn in your geek card immediately or risk the consequences.

Re:That 0.02 TB made the difference.

[halltk1983]

*OR* manufacturers of SSDs overprovision them and then use the extra capacity for write leveling, but I'm sure you're probably more right than they are.

Re:That 0.02 TB made the difference.

[EvilSS]

MFM or RLL?

Re:That 0.02 TB made the difference.

[haruchai]

I'll take the 30 TB, you can have the slice of pi

Re:That 0.02 TB made the difference.

[Unnamed Chickenheart]

Wooooâ"haa!

Why stop at 11 digits? You need to think of the neglected digits after the dot!

Re:That 0.02 TB made the difference.

[fisted]

Who (and why?) produces NAND flash chips that don't have a size in bits which is a power of two? That doesn't make sense.

Re:That 0.02 TB made the difference.

[TheRaven64]

Your co-worker was probably not thinking very clearly. Back then, floppy disks were 360KB and so 10MB was under 30 floppy disks worth of capacity. A few years later, floppy disks were up to 1.44KB and stayed there, while hard disks grew to 500MB or more. At that size, it looks hard to fill. Then CDs came out and could store the equivalent of 450 floppy disks and that seemed a huge amount. A 2GB hard disk could store 3 CDs and still have space!

Re:That 0.02 TB made the difference.

[sconeu]

MFM. ST-251, instead of the stock ST-225.

Re:That 0.02 TB made the difference.

[EvilSS]

That was my first hard drive. Bought it used as a kid and was in hog heaven since all I had before that was dual 5.25 floppies.

Re:That 0.02 TB made the difference.

[sconeu]

Since I had a stand for the machine, I even had the vendor run the low-level format with the drive vertical.

For you youngsters, back in the day, there were no tower cases, you stood a desktop case on end. In addition, you had to run a "low level format" to physically mark the sectors on the drive, as well as mark the sector interleave. If you were running the drive vertically, instead of horizontally, you got better performance when you low leveled it in the vertical orientation.

New metric

[Anonymous Coward]

Instead of LOC, it should be LOP (Library of Pr0n). So how many of these drive will it take to store a single LOP?

Re:New metric

[HornWumpus]

At 1700 MB/s the right question is: How many will it take to keep up with the current growth rate of the LOP.

Re:New metric

[Anonymous Coward]

Good point! At the current delta growth rate, that could far exceed the bandwidth capabilities to keep up.

Re:New metric

[DontBeAMoran]

In other words, they're fucked.

Re:New metric

[rogoshen1]

eventually with the deepfakes type stuff, we won't need LOP's anymore, just a program to generate on the fly whatever midget+monkey+donkey fetish the mind can conjure up.

Coupled with VR, this will spell the end of mankind -- constant, ever increasing levels of titillation and depravity.

Only $25,000

[110010001000]

I got a quote for these. Only $28k. Pretty good job by Samsung. The 16TB previous gen are about $11k.

Re:Only $25,000

[110010001000]

Ooops, $28k, not $25k. But what is $3k worth anyway? I can't even get a bitcoin for that!

Re:Only $25,000

[jellomizer]

At $0.81 a Gigabyte that is still a lot. Especially being traditional drives are between $0.02 and $0.05 a Gigabyte

I hope these disks are more reliable then a Raid-5

Re:Only $25,000

[EvilSS]

Why would they need to be more reliable than a RAID 5? These aren't going into anything as single drives but as part of large dedicated storage systems. Also who the heck is still doing RAID 5 at the enterprise level? What is this 2002?

Re: Only $25,000

[c6gunner]

Either that was an unnecessary correction, or you need a correction to your correction ...

Re:Only $25,000

[ctilsie242]

I would assert they will be used in the usual RAID 1/10/6 configurations. They are fast, but because there is so much data on them and I/O hasn't kept up, RAID 5 would likely not be used, but would be used for RAID 6. Of course, RAID 1 and RAID 10 will definitely be alternatives.

I do wonder what the real world failure rate on these will be. From what I've seen, SSDs have definitely been more reliable than HDDs.

Re:Only $25,000

[jabuzz]

Any competent storage administrators who understands about mean time to data loss (MTTDL) knows that there are still use cases even in 2018 where RAID5 makes sense. If my MTTDL is 20,000 years using a RAID5 why would I need to move to RAID6?

Sure if I am building a files system of a few hundred TB or more made of dozens of eight disk plus parity RAID arrays then RAID5 is inadequate. The problem is ignorant people have seen/heard storage administrators who make such petabyte scale file systems talk about RAID5 being inadequate and presumed without understanding the issue that it applies to all configurations and use cases of RAID5.

Further as an administrator of petabyte scale file systems I can tell you now that even RAID6 is out the window. The rebuild times are just too awful, it's all distributed RAID/dynamic disk pools (no consistent naming for the technology) these days.

Anyway need 4TB of RAID backed SSD, then 2D+P RAID5 made out of 2TB disks is a lot cheaper than a RAID1 made of 4TB disks and actually has a larger MTTDL so is better. I guess you could do a 4D+2P RAID6 of 1TB SSD's but you are burning a lot of drive slots their and you have more points of failure, and the MTTDL is not significantly better. You might go from 20,000 years to 25,000 years so who gives a dam. What you need to avoid is a MTTDL of 2 years or even 10 years. Frankly a MTTDL of 100 years makes me nervous, but 20,000 or 25,000 is all the same because the dam thing is backed up. If I *REALLY* care then you would use DIF/DIX that pushes it out way further.

Re:Only $25,000

[EvilSS]

Pretty much every storage vendor has gone to some flavor of DP or even TP, with RAID 0 or some other proprietary scheme to span shelves. Between that and hot spares drive failures become a non-urgent issue for the customer and hardware vendor. Drive fails, the system phones home, drive and (depending on your service contract) vendor show a day or two later to swap it out. The added cost is usually not much compared to the overall cost of the system. We also see RAID 4*, mainly on smaller SSD (14 disk shelves) deployments but very rarely RAID 5 anymore. RAID 5 has just fallen out of favor since most places are not doing server RAID arrays much anymore, but using SAN/NAS systems for data storage instead. It makes sense if you are limited to 5 or 6 drives in a system but even SMBs are buying 14 and 28 disk shelves and consolidating storage away from individual servers due to the better backup/recovery options and overall ROI these systems offer.

*The vendors like RAID 4 over 5 because although there is some performance hit keeping the parity data on one disk, there is no performance degradation when the array is in recovery mode. RAID 5 runs a bit faster in normal ops, but can degrade quite a bit when a drive is lost/rebuilding.

Now if this disn't cost more...

[cdu13a]

then a new station wagon I could upgrade from all those damn tapes.

I have to wonder

[DaMattster]

How much is this behemouth going to cost? It cannot be cheap.

Re:I have to wonder

[angel'o'sphere]

A car and a leg!
Or two legs, even!

why not pci-e?

[Joe_Dragon]

why not pci-e?

Re:why not pci-e?

[the_skywise]

For one it's probably intended for cloud storage services so they intend to hook a LOT of them up to one storage cluster which generally uses SCSI.
Secondarily, faster transfers means more heat and they may be forcing the SCSI interfaces to "throttle" the transfers and keep the thermal issues manageable.
I was surprised how easily thermal issues could pop up on just a 512gb M2 stick.

Re: RAM

[Anonymous Coward]

RAM is far faster than flash, and far more tolerant to being written lots. Itâ(TM)s used as a cache to both speed up the drive, and reduce wear on the actual flash.

Beowulf cluster....

[Anonymous Coward]

...of computers with these? you gotta wonder.

RPM of drive?

[Anonymous Coward]

Is this a slow 5200 RPM drive, or does it actually spin at 7200 RPM or even enterprise-level 15k rpm?

Re:RPM of drive?

[Freultwah]

As SSD means "Super Spinning Disk", it probably goes to at least 30k.

Re:RPM of drive?

[DontBeAMoran]

Virtual +1 Funny.

Please use proper capacity units.

[sconeu]

30.72 TB is approximately 2 Library of Congresses.

https://blogs.loc.gov/thesignal/2012/04/a-library-of-congress-worth-of-data-its-all-in-how-you-define-it/

Re:Please use proper capacity units.

[Spaham]

Yeah but they don't store porn...

Re:Please use proper capacity units.

[EvilSS]

OK but what is that in bananas?

I was wondering about this

[Krishnoid]

Since Samsung was the one who engineered stackable V-NAND, I suspected that they could multiply the current market's SSD capacities by the layer-count. Maybe that's not what's happening here, but it sure seems like Samsung will have the edge in increasing SSD capacities.

Re:Just rolls off the tongue

[WilliamGeorge]

"PM1643, why aren't you at your post? PM1643, do you copy?"

Re:Meanwhile, in 2016....

[DontBeAMoran]

So, time is going backwards?

Time to sell all my Bitcoins right now and buy them at 8 cents each in 2010!

Monster pictures

[VeryFluffyBunny]

Great! I've been looking for something to store my enormous collection of monster pictures on :)

Re:Why not a 3.5'' ssd drive

[omnichad]

This one looks to be. That's how big you have to go to need such a large form factor.

iscsi sucks (Ceph man) wait ESXI can't use that wi

[Joe_Dragon]

iscsi sucks (Ceph man) wait ESXI can't use that with out an iscsi gateway.

Form Factor?

[kenh]

I'd love to see this in a Full-Height 5 1/4" form factor for my old Pentium P133 server...

Sadly we're still stagnated

[AbRASiON]

Around three years ago, we had articles claiming that we would have 128GB SSDs by this year.

The market went backwards when that price issue hit around near 2 years ago.

All I'd like (and have wanted for 3 years) is a "slow crappy" SSD that's "only" say 400MB/s sustained read and write, with typical SSD access times (or no more than half as slow as current ones)
However that being said, in the 6 to 10TB range under 500$ USD.

Looks like it's going to be 2026 at this rate unfortunately.

BAH!

[gabrieltss]

My 320 MB SCSI Maxtor hard drive I bought in 1987 is still a live and kicking. and only half full. Oh wait that's in my Amiga 2000 - that's why it's only half full - no bloated software!

But Mr. Gates told us "No one will ever need more than 640K." so that means you don't need a 30+TB SSD - you only need a 12 MB Corvis Hard Drive!

Re:BAH!

[gabrieltss]

Sorry but your wrong - I have the drive in my possession. I'd post a pic of it but no way to attach to this post.

flash shortage much?

[slashmydots]

Hey guys, you know about that flash chip shortage? I found them. Here they are.

Re:Evidence to show these drives push down prices?

[EvilSS]

You want evidence of SSD prices dropping over time?

Re:About those read/write speeds...

[Khyber]

DING DING DING! We have the winner!

As soon as I saw the connection I knew those speeds were utter bullshit.

amd epyc has the pci-e io to drive a lot of disks

[Joe_Dragon]

amd epyc has the pci-e io to drive a lot of disks with just 1 cpu. Just 9 disks at X4 + 2X 10 gig-e maxes out 1 Intel cpu.

small sata dom's also work for boot even on storag

[Joe_Dragon]

small sata dom's also work for boot even on storage severs 8GB is fine for Linux unless you really a lot of room for logs.

AMD EPYC is for you PCI-E no switches needed.

[Joe_Dragon]

AMD EPYC is for you PCI-E no switches needed.

Do they tell you....

[OppMan29]

How many pictures ... songs and HD video hours it can hold...

ceph and HBA in IT mode works and easy to move

[Joe_Dragon]

ceph and HBA in IT mode works and easy to move disks if an server dies (also your storage will stay up if an full server goes down)

Yes do that os update needs an reboot live.

Re:RAM

[TheRaven64]

RAM is byte addressable. Or, practically, cache-line addressable. You read and write to DRAM in chunks of 64 bytes on most modern processors. Flash is byte-writeable but only cell-erasable. If you want to erase and overwrite, you have to do so in chunks of the cell size, which is typically a multiple of the exposed block size (4KB on modern machines). It would be possible to design a cache controller that always flushed 8KB chunks back to main memory, but then it would be really slow. That said, you could more easily design a fourth-level cache that worked at a page granularity and used DRAM so you'd flush from DRAM to SSD in some multiple of page granularity. That's more or less what HP's The Machine is doing.

Usable space

[tgrigsby]

Sadly, Windows will only recognize the first 3.5 GBs....