How long will HDD's remain relevant to home users?

llmercll

[H]ard|Gawd
Joined
Feb 17, 2010
Messages
1,369
Hi everyone,

I've noticed that within the past 2 years SSD's have become significantly larger and cheaper. It wasn't long ago that a 120GB SSD was considered large, and was quite expensive. Now, they are more comparable to 512GB SSD's. in price anyway.

HDD's have been getting larger as well, and I believe 12TB capacities exist, but more affordable and realistic for home users is a 4TB, or maybe 6TB. A few years back, 2TB was considered large.

I've been thinking lately about when, and if, SSD's will overtake HDD's as the primary form of storage. SSD's are lighter and smaller. There's no moving parts either. I am not sure if they use less energy, are more reliable, or have a longer MTBF, however.

I imagine as long as there are hard drives, they will be cheaper $/GB than SSD, and therefore more attractive to consumers. I don't really know why they will remain cheaper, however. It seems price would be the only reason for them to remain relevant to home users (and I suppose large data centers as well). But what, hypothetically, is even keeping manufacturers from stacking a bunch of NAND memory into the dimensions of a 3.5" HDD. If they can fit 1TB into the tiny 2.5" drives, what's stopping them from putting all that memory into a 3.5" drive and getting a 4TB or higher SSD? I'm not sure why this hasn't been done yet, but I am sure there is some technical limitation I'm not aware of.

I feel like as a % of GB increase, SSD has grown more rapidly over the past few years. While HDD's have gained more GB, SSD has gained a larger increase/$.

Most of what I've read on the internet, on forums and such, posters argue that SSD won't overtake HDD for a LOOONG time. I haven't seen many explain why though. I am curious to the physical limitations on how dense HDD platters can get. Are we reaching their peak density? I've noticed some manufacturers are using helium now. Can the SSD theoretically be made denser than the HDD? This is probably a bad analogy, but HDD seems more "analog", while ssd seems more "digital" to me.

Can someone explain this for me?

Thank you!
 
On the HDD side, there's a lot of ostriches sticking their head in the sand. Seagate is the biggest ostrich of the bunch (see: their official position in 2012). WD just got their head out of the sand and bought Sandisk, leaving Seagate having to figure out how to buy Micron or build their own NAND flash factories before NAND flash becomes irrelevant.

Yeah, that last part. That's when the ostriches get eaten.
 
I think part of this will be perception. Some people will always argue the numbers: more capacity, lower price, faster speed, whatever metric you want to pick. That could point to any number of different devices, depending on what you choose as your primary metric.

Personally, I think that "Storage" has hit the "big enough" stage for a while now. Sure, there are a lot of people who use a lot of TB of data - but how much of that do you keep on your tablet, or your laptop, or the PC itself - without shuffling it off to a dedicated storage device once you start getting that large?

Once you hit "big enough" - which may be a bit different for different people, but I loosely define as 1TB based on the people and clientele I've worked with in my area - then your really just looking at price and speed. Some people are more sensitive to the price, some more sensitive to the speed. I tend to recommend SSD when I can make it fit comfortably in the budget (or can safely downsize on the capacity). I often recommend a dual drive approach, a small SSD for the boot volume and a bulk HDD for pictures/videos/whatever storage capacity. I like Apple's Fusion drive and wish I had something comparable on the PC end that was as easy to use, but I haven't found one yet.

So.. yeah, I don't think HDDs will die off for a long time in laptops or PCs. There will always be some people who just look at the sticker and say "4TB > 1TB, who cares if it's SSD or HDD", but I think anyone who's used an SSD-primary drive computer will have a hard time going back to an HDD. And if you think about it, I haven't seen an HDD in a "mobile" device since the iPod Classic died off in what, '09? Everything since has been some form of non-volatile memory (albeit not SSD exactly)

I think that SATA will be the next thing to go - once we see that go away (maybe it'll be replaced by NVMe, maybe something else, hard to say yet), we'll be one or two generations away from the HDD going away in the consumer space, since I don't see HDDs making the jump to the next storage interface format.

What is stopping them from making 3.5" SSDs? They do make them. They mostly exist in the enterprise space. They are expensive. And current controllers can only address a finite number of channels (basically the number of NAND chips they can talk to at once) - not to say you couldn't redesign one with more channels or whatever, but there are some technical limitations to that answer as well beyond just what the price would be.
 
On the HDD side, there's a lot of ostriches sticking their head in the sand. Seagate is the biggest ostrich of the bunch (see: their official position in 2012). WD just got their head out of the sand and bought Sandisk, leaving Seagate having to figure out how to buy Micron or build their own NAND flash factories before NAND flash becomes irrelevant.

Yeah, that last part. That's when the ostriches get eaten.

Seagate bought LSI Flash in mid '14.

The same LSI that bought SandForce in '11. And there are a lot of SandForce-based drives out there.
 
HDD's will be relevant until the main manufacturers finally transition their factories to NAND factories. We will probably not see a large jump again on spinners.
 
I was explaining this to a friend of mine:

SSDs have doubled in both size AND value in the last two years, wheras HDDs have added roughly ~20-50% of space in that time and prices have yet to really move much.

Essentially: SSDs are increasing in capacity and getting cheaper MUCH faster than HDDs.

HDDs may be cheaper per GB now, and for a few years: but make no mistake: SSDs WILL become cheaper per GB than HDDs within the decade, at that point: HDDs will become the new Tape Drive: around, but not relevant.
 
Right now, if you need a lot of space, SSD is still pretty expensive, but if you don't need too much, ~$100 of SSD gets you probably good enough. For large storage, hard drives are still good; although capacities are somewhat stagnant, 4tb is a lot of storage for most people.
 
They're good for mass storage, file servers, and surveillance recording, but they're obsolete as OS drives.
 
HDDs may be cheaper per GB now, and for a few years: but make no mistake: SSDs WILL become cheaper per GB than HDDs within the decade, at that point: HDDs will become the new Tape Drive: around, but not relevant.

That is a risky prediction to make. It could turn out to be true, but if HDDs manage to make another breakthrough in density in this decade (with, say, HAMR or patterned platters), then it could be wrong.

Today, SSDs are about 12 times more expensive per TB than HDDs. If HDDs do not improve at all in $/TB over the next 10 years, then SSDs would have to improve by 12 times, which would be a doubling in TB/$ every 2.7 years. That seems feasible.

But if HDDs continue to improve in TB/$, lets say doubling every 4 years, then the SSDs would need to improve by 12 times multiplied by 5.6 times = 67 times, which in 10 years comes to doubling every 20 months. That may be feasible with 3D flash, but it remains to be seen.

But if HDDs continue to improve by doubling TB/$ every 3 years, then they would improve by a factor of 10 in 10 years. SSDs would need to improve by a factor of 120 in 10 years to catch the HDDs. That would be doubling every 17 months, which if I had to bet, I would take the over (i.e., SSDs will have a longer doubling time than 17 months over the next 10 years).

Finally, if HDDs manage to double TB/$ every 2 years, then they will improve by 32 times in 10 years. So SSDs need 384 times to catch HDDs. That comes to a doubling time of 14 months. I consider that unlikely.

Bottom line is that I think saying that SSDs will definitely catch HDDs in the next 10 years is a bit optimistic. If you had put the over/under at 10 years, then I think you would have a good bet -- the likelihood that it will take less than 10 years is about the same as that it will take more than 10 years.
 
That is a risky prediction to make. It could turn out to be true, but if HDDs manage to make another breakthrough in density in this decade (with, say, HAMR or patterned platters), then it could be wrong.

Today, SSDs are about 12 times more expensive per TB than HDDs. If HDDs do not improve at all in $/TB over the next 10 years, then SSDs would have to improve by 12 times, which would be a doubling in TB/$ every 2.7 years. That seems feasible

Your only looking at the $/byte metric. There is a performance metric that should be considered as well.

Personally, I'll take 10x faster over 10x bigger - so long as either is "big enough" in the first place.

Catching HDDs - there's a lot of different ways to look at that, and everyone gets hung on $/byte and ignores everything else.
 
Personally, I'll take 10x faster over 10x bigger - so long as either is "big enough" in the first place.

That is a totally different question, and not answerable because "big enough" is not well defined. Also, you are ignoring costs. A lot of people will choose cheaper over faster, so long as it is "fast enough". :p
 
HDDs may be cheaper per GB now, and for a few years: but make no mistake: SSDs WILL become cheaper per GB than HDDs within the decade, at that point: HDDs will become the new Tape Drive: around, but not relevant.

Related to what JoeComp said above, I'm not sure the price gap is going to close that fast. Ignoring the obvious speed differences, in terms of just capacity, I remember doing a pricing analysis about a decade ago. If I recall, SSDs at the time were about 16 times more expensive per capacity than HDDs. I was responding to someone over on NBF that made a similar claim about SSDs overtaking HDDs in price per GB in probably a similar time frame of about a decade. I forget the exact numbers, but I was looking at the rate at which HDDs were increasing in capacity and decreasing in price as compared to the (fairly limited) pricing data of SSDs at the time, and there was no way they were going to even catch up in price in such a "short" period. Here we are, a decade later, and that gap has closed only by about half (maybe closer to a quarter or third, if you're looking only at 2.5" drive pricing).

Your only looking at the $/byte metric.

...which is exactly what the quoted post claimed, that SSDs will be cheaper per GB than HDDs within the decade.
 
You won't have to worry about SSDs simply being 'Big enough', as no doubt Samsung will have a 3 and 4 TB SSD in consumer's hands at around the $1K mark in th next 24 months, 8TB within 5 years for the same price. HDDs will barely have broken 10 TB at that point. The demand will fall onto SSDs as the PRIMARY source of data storage. With HDDs simply becoming the 'cheaper' option initially. Not faster, not larger, not more efficient, not more reliable... just cheaper. The demand will drop off and SSD manufacturers will eventually compete with each-other down to the 10c per gigabyte level. In 10 years, HDDs will be irrelevant.
 
You won't have to worry about SSDs simply being 'Big enough', as no doubt Samsung will have a 3 and 4 TB SSD in consumer's hands at around the $1K mark in th next 24 months

How about in 8 months for $800 instead?

The Samsung 850 Pro 4TB comes out in 2016. That will probably happen in June or July, because Samsung has been making major SSD releases in that part of the year; that's when they released the 2TBs.

Also, the price of the 850 Pro 2TB has already dropped from $1000 to $860. The 2TB 850 EVO launched at $800, is $740 and was on sale for $650 once. The $1000 and $800 slots will certainly be wide open in another 8 months.
 
How about in 8 months for $800 instead?

The Samsung 850 Pro 4TB comes out in 2016. That will probably happen in June or July, because Samsung has been making major SSD releases in that part of the year; that's when they released the 2TBs.

Also, the price of the 850 Pro 2TB has already dropped from $1000 to $860. The 2TB 850 EVO launched at $800, is $740 and was on sale for $650 once. The $1000 and $800 slots will certainly be wide open in another 8 months.

So essentially, SSDs (at least Samsung SSDs) are doubling in $/byte every year. So we may expect to see >HDD levels of capacity in consumer 2.5" SSDs within the next 5 years.
 
How about in 8 months for $800 instead?

The Samsung 850 Pro 4TB comes out in 2016. That will probably happen in June or July, because Samsung has been making major SSD releases in that part of the year; that's when they released the 2TBs.

Also, the price of the 850 Pro 2TB has already dropped from $1000 to $860. The 2TB 850 EVO launched at $800, is $740 and was on sale for $650 once. The $1000 and $800 slots will certainly be wide open in another 8 months.

oh man i so want them spinners to be gone, but with 4tb REs at $250 i guess i'll still be waiting.
 
To be honest, I don't use TONS of data, and all my HDDs are in a server. my main desktop (not in sig, new) only has 2 500GB SSDs. 1Tb SSD space with HDDs on a server. If I were like some of me customers who filled up a 2TB HDD every month with pirated movies, I may think differently. But as far as my perception is concerned: There is no room for HDDs in my personal computers from now on. Servers? maybe.
 
You won't have to worry about SSDs simply being 'Big enough', as no doubt Samsung will have a 3 and 4 TB SSD in consumer's hands at around the $1K mark in th next 24 months, 8TB within 5 years for the same price. HDDs will barely have broken 10 TB at that point.

You think HDDs will only be at 10TB in 5 years? I'd certainly bet against you on that.
 
I think several people here are overly optimistic on the increase in SSD capacities and TB/$ , and overly pessimistic about the same on HDDs.

No doubt SSDs will grow more quickly than HDDs in the next decade.

But I think one factor that will slow the growth of consumer SSDs somewhat is that, as has already been mentioned in this thread, few consumers really need a drive (SSD or HDD) larger than about 1 or 2TB.

Now, you could say the same about consumers and HDDs, except for one thing. There is a niche market in consumer HDDs that is all about capacity. Mostly it is for people who are storing a lot of video and movies on their home servers. And that niche is very cost sensitive. Until SSDs actually match the price of HDDs per TB, those people will continue to buy HDDs.

So, on the one hand you have some sort of niche market for consumer SSDs larger than 2TB, and on the other hand you have a niche market for consumer HDDs as large as possible. Until the SSD price per TB matches HDDs, my bet would be on that HDD niche market over the SSD niche market. It is a tough call. They are both small, niche markets. But at least I can think of a use some consumers might have for 20TB HDDs (especially if 4K movies become common). I cannot really think of what use most consumers would have for a 8 or 16TB SSD. So even if SSDs could match HDD prices in 10 years, how are they going to get there in the meantime? Who is going to be buying those 8TB SSDs in 5 years, when they are still several times the price of an equivalent HDD? Maybe there will be a business market for that size of SSD, but then the prices are going to be different, no longer consumer level.
 
Last edited:
Basically hard drives will remain relevant as the default for bulk storage devices. Even now, flash devices can't beat their price-per-meg/gig/tera.

However, there's probably going to be a squeeze in the HDD market with regards to reliability eventually.

Sure, you can get a 4/8/12/whatever terabyte drive. But if it falls over and dies in 2 years, what good is it?
 
until I can get a 2TB SSD for <$50 I would not consider them useful for bulk storage
 
Huh? If you have a good backup plan an HDD failure should never wipe you out.

I don't see that he mentioned anything about data loss, only drive failure. Drive failure is still a pain in the ass even if you don't lose any data. You can always add more spare drives, or go double parity, etc, but if you're spending so much extra to compensate for the poor reliability of the drives, at what point do you factor that cost into the initial purchase decision?
 
Huh? If you have a good backup plan an HDD failure should never wipe you out.

Not addressing data loss.

I'm talking, in general, about cheapshit hard drives that pretty much are designed to fail almost the second their warranty expires.
And how, with SSDs coming along, that hard drive manufacturers should be focusing on quality and endurance for their future drives.
 
Hard drive development has become glacial since the Hitachi 4Tb came out years ago.

They used to double in size every year.

SSD's are already at 2Tb and mechanicals are still stuck at 6Tb.
When SSD's were 128Gb, mechanicals were already 3Tb.
 
Don't forget that there was a major disaster that basically HALTED most hard drive production several years ago.
 
For as long as people need to store large amounts of data. However, not a huge percentage of home users have ever had that need. And with an increasing amount of entertainment media delivered through streaming services, that percentage becomes smaller every day. Increased broadband speeds also mean a greater ability to rely on online storage for everyday needs like backups and photo and music storage.
 
You think HDDs will only be at 10TB in 5 years? I'd certainly bet against you on that.

I may be wrong, but the way I see it, HDDs have hardly increased in size in the last 24 months. The biggest new tech was shingled datacells and those slowed reads down like crazy... and barely increased the maximum capacity.

At the rate we are seeing now, SSDs will completely surpass HDDs with regards to maximum consumer capacity in a few years. Its highly likely you will see 24TB SSDs on the same shelf as a 16TB HDD, and the year after that, a 48TB SSD on the same shelf as a 20TB HDD.

Yeah, the HDDs will remain cheaper for a little while, but the mass-storage whores will have a bit of a conundrum when they stare down a 2.5" drive that can hold twice as much as the single largest capacity HDD on the market.
 
Don't want to fork out for NAND flash? You're not alone. Disk still rules

http://www.theregister.co.uk/2015/10/05/flash_shipments_not_catching_up_with_disk/

pb_shipped_flash_and_disk.jpg


Why isn't shipped flash capacity catching up with that of disk? Cost is a large reason. Rakers said Samsung is estimated to be spending over $23bn in capex on its 3D NAND for for an estimated ~10-12 exabytes of capacity.

Say, on this metric and crudely, every 10,000PB of NAND capacity costs $20bn, then to catch up with HDD capacity shipped in 2019, the flash industry would have to spend $2tn. We don't think it is going to happen unless flash capacity $/GB leaving the foundry is sustainably lower than that of disk.

They did not get into the details in the article, but the basic issue is that yields for 3D flash have not been high. While increasing the number of layers provides a clear path to increasing flash storage density, the investment required, in capex and process development, is very high. Unless there is a huge expansion in demand for flash memory, then the industry will keep the yearly capital expenditures low enough to turn a profit. That means slower growth in flash storage.
 
For as long as HDD is cheaper than SSD per GB. I just can't see myself using SSD for backups when HDD is cheaper.
 
HDDs are quite handy for video editing & recording and storing generic data (pictures, music, videos). Two modern HDDs in RAID0 are really close to standard SSD in sequential read and write speeds. HDDs will stay relevant until SSDs are cheap enough and offer really good endurance (so that you could use cheap ones in PVRs etc. without having to fear about it wearing out).
 
I'm ok with hard drives being completely driven out of the desktop/individual computer market, the sooner the better!

They are already almost completely gone from mobile, only the very cheap laptops bother with spinning rust anymore and even that is shrinking fast. (not counting the DTRs that are barely a blip on the radar)

Hard drives will stick it out in the enterprise network/cloud/etc bulk storage markets for a long time, as expected. Those types of customers will generate plenty of demand to keep the HDD vendors (all 3? of them) chugging along, they are already the biggest part of the market anyways.
 
For large bulk storage like NAS still relevant since 10TB are available with 6TB being more mainstream right now.
For C drive a no-brainer to go SSD especially if you have SATAIII, M.2, etc.

Still hoping a future RST allows complete pass through for raid1 and SSD so that programs can "see" the array so you can get SMART attributes in real time.
"
 
Hard drives can be written to ad nauseum and that will be important for many things where writes are intensive.

I used to think SSD's would hit a brick wall in reliability (and still do) as the NAND gets smaller, but this is somewhat happening to hard drives too. Reliability of multi-terabyte HDD's have been abysmal.
 
Hard drives can be written to ad nauseum and that will be important for many things where writes are intensive.

I used to think SSD's would hit a brick wall in reliability (and still do) as the NAND gets smaller...

Neither of those are serious concerns.

A 2TB SSD should be good for at least 1000 full drive writes, which is 2,000TB. That comes to 22GB per hour for 10 years. PER HOUR. Virtually no reasonable workloads are going to exceed that, or even come close.

As for SSD endurance decreasing with lithography size, that is not a concern for quite a while now that the industry is rapidly switching to 3D flash, where the lithography sizes have actually taken a leap to much bigger features.
 
Most home users do not need more than 256GB, if the sub-1TB SSD prices can match or beat the sub-1TB HDD prices HDD's will no longer be relevant for 99% of home users.
 
most people I know don't need more than 256GB of storage so all of them could go SSD.
I have 18TB on my media server and am running out of space, but have no funds to upgrade the 7 2TB drives to larger ones.
 
I have 18TB on my media server and am running out of space, but have no funds to upgrade the 7 2TB drives to larger ones.

That is a good use case for SnapRAID. I mean for my media usage on my linux based pvr I mix and match drive sizes. Today I purchased 2 x 5TB drives from a forum member for $130 each these will fit fine with the mix of 8 x 2TB to 4 TB drives I currently have. Although my 2 external parity disks are only 4TB so I will have to use the extra 1TB for something else (which is no problem since the linux box is also my main server).
 
Last edited:
I may be wrong, but the way I see it, HDDs have hardly increased in size in the last 24 months. The biggest new tech was shingled datacells and those slowed reads down like crazy... and barely increased the maximum capacity.

At the rate we are seeing now, SSDs will completely surpass HDDs with regards to maximum consumer capacity in a few years. Its highly likely you will see 24TB SSDs on the same shelf as a 16TB HDD, and the year after that, a 48TB SSD on the same shelf as a 20TB HDD.

Yeah, the HDDs will remain cheaper for a little while, but the mass-storage whores will have a bit of a conundrum when they stare down a 2.5" drive that can hold twice as much as the single largest capacity HDD on the market.

While the areal density of magnetic drives is definitely petering out, using present SSD density trends as a future predictor seems, um, optimistic. They were, for quite a long time, process-wise way behind the ball, but a huge portion of the gains has been in pushing down to the lithographic nodes associated with other chips (and picking up all kinds of loose hanging fruit). And as much as we've staved off Moore's law this far, it honestly appears that we're running out of atoms. Continued 3D fabrication will certainly benefit NAND flash, don't get me wrong, but we're very much in the golden era of SSD improvements and future trends will start bending down.

As for the consumer space, I see NAND taking over more and more, especially in laptops. They're smaller and lower power--not to mention shock resistant. I've certainly been noting that generation after generation of laptop is less bulky than the last. :)
 
Back
Top