That cost did not go up much compared to amount of extra fired people, was kinda on the horizon that this would happen, good luck to them. Did WD start to cut already, I mean they have SANDisk, HGST,.... so many things still merging?
So apparently, even cutting warranty down to the laughable 12 months their business model still sucks, thus the need to kick out people who have nothing to do with it, while the lousy management lingers on...
For being too slow to realize that SSDs were going to disrupt the HDD market and not moving into the new business fast enough, yes I do blame Seagate's management. Along with WD they're hemorrhaging HDDs sales and not bringing in nearly enough SSD to replace them. I think Toshiba is better off here; but that has as much to do with being a bit player in the HDD market as to being a major player in the flash one.
Completely agree here. They simply didn't invest enough in SSD. For that, you can definitely blame mgmt. Five years ago, they should have seen what was coming, and started their investing then. They could have been co-leaders in the retail/consumer SSD market along with Samsung.
Of those 8,000 jobs being cut, how many are the higher-up ones who lacked the foresight & also didn't want to change direction when the tide was clearly turning?
I say five years ago, but I guess they should have started investing serious money as soon as they saw the mainstream Intel X25-M doing well in 2008. That alone should have told them where the future was going.
Sandforce, though, provided a moderate performance platform so the only thing left over was the really low end and the higher end. Even Intel ended up producing Sandforce drives. I'm not sure what Seagate would have been able to do, other than try to beat Samsung to the market. The 830 series were good drives. And the lackluster 840 TLC stuff was massively hyped and was a major market success.
I suppose Seagate could have bought Sandforce instead of having them be closely tied to OCZ (the result of which was highly unstable drives like the 64-bit NAND Vertex 2 240 sold with false specs on the packaging) — and a horrible "panic mode" designed to destroy the ability of a user to get their data so that Sandforce would sleep peacefully knowing no one would be able to pry into its subpar firmware that doesn't work with TRIM correctly.
Well, as a counterpoint, the current market leader - Samsung - uses their own controllers. They didn't resort to Sandforce. Even if Seagate didn't want to partner with SandForce, they could have partnered with Marvell to use their controllers while they were building their own internal expertise in controllers. Instead, they did neither and mainly tried to stay status quo with HDDs.
Meanwhile, SSD prices were going down and performance/capacity was going up. Seagate stays head-in-sand and still doesn't try to get a complete SSD product line. Samsung fills out their product line with TLC-based drives along with already well-performing MLC drives, and pretty much becomes the market leader. Even WD buys into their own future by acquiring Sandisk, thus ensuring they'll still have life after HDDs really tank. I think Seagate should be looking at acquiring Kingston or Micron to keep their future vibrant rather than just a slow decline.
Seagate bought SandForce from Avago in 2014, so they currently have in house controller just as Samsung.
The only viable option for Seagate is to aquire Micron or SK Hynix. Buying Kingston wouldn't add any value to Seagate, as they can just as well buy the nand from the market and package it.
Going after Micron as a whole probably wouldn't make sense; too big (about 50% higher market cap than Seagate ~13bn vs 9bn) and has too many other business units that wouldn't be of any interest (eg dram). Since the flash production is via a joint subsidy they co-own with Intel it's possible the latter might have a poison pill in place to protect its share of the investment.
SK Hynix is even bigger (~ 22bn USD). Being part of the 3rd largest conglomerate in South Korea the barriers to any sort of takeover are probably insurmountable. (South Korean holding structures tend to look like Rube Goldberg machines and are designed to let the founders have majority voting rights even though they own a much smaller fraction of the total shares on the market.)
Yep, that was a great concept to cache high io hits, independent of the file system. But for some reason they never really perfected it with more NAND and enabling write caching.
SSDs are a major part of it, no doubt, but you should also look at what gave SSDs such a large opening to start killing off their business so soon. I'm talking about the price-fixing in the wake of the 2011 Thailand floods. The HDD manufacturers enjoyed a turn around from the price war they had been fighting and artificially kept the prices elevated when they should have dropped after the flood. IMHO today they still haven't returned to the rapid price drops they had before. People being pushed to pay more changed 3 behaviors: 1. People moved to cloud storage to offload the risk of HDD prices, and cloud providers could consolidate the number of disks used in a way many individual users couldn't. 2. With the price premium for SSDs dropped, people who could afford it started moving to them for their main drives instead of HDDs 3. End users kept disks in use longer then they would have before and brought older disks that had been retired back into service to fill the gaps.
All this reduced the market for HDDs and the manufacturers are now paying the price. Add in the collapse of the PC market as equipment cycles were extended across the board, and you can see why they are in trouble.
In addition to cloud storage, cloud providers (streaming and document storage, Office 365, etc) killed the need to store a lot data locally. Enterprises are moving a lot of infrastructure to the cloud.
According to articles Anton's written about the HDD market, low end consumer drives do return a very small return above manufacturing costs.
But it's the expensive max capacity and enterprise models that generate the large marginal profits needed to fund R&D; which's why the slump in $50 laptop sales looks a lot worse from sales volume numbers than it is from a profitability/R&D impact perspective. OTOH when SSDs manage to gut the 10/15k SAS market; that will hurt their ability to sustain R&D a lot more. (And with Seagate thinking its next 15k generation might be the last at that speed the beginning of the end might be near.)
No real improvement unfortunately. There's just not a scenario where it makes sense. If the SSD side were bigger and you could choose what you wanted to store on there, it might work.
They've more or less withered and died. The original models were things like 4/256GB. The 4GB of flash was too small to be useful for much beyond a write cache; and even there the simplistic write tens of gigabytes of data testing that was common at the time meant that it wasn't benchmark visible. At the time a 32/64GB flash 500gb laptop HDD (or 2/4TB desktop HDD) combo might've offered enough flash to hit the sweet spot if coupled with software/drivers smart enough to cache OS/Application binaries and frequently accessed user data on flash while archiving apps that haven't been used for months and media to the HDD part. Only Apple actually delivered a hybrid drive with that sort of resource mix; never having used one I can't comment on how well their software delivered.
Now the problem for relevancy is that an SSD big enough to be suitable for most users is about the same cost or cheaper than a smaller but still useful SSD and a small laptop HDD. The only groups who still might need something like that would be gamers on a budget and people with large media collections they don't want to offload to a NAS. The problems there are both that those are relatively small groups of users, and that they're also the users least likely to find having to manage multiple drives a serious burden.
Intel has offered ways to pair SSD/HDD in that way, with up to 64GB of SSD caching most used data automatically in your HDD. I've been using this solution for years (this came our of the z68 chipset, core i 2xxx generation). This works ridiculously well with only 32GB for the cache.
So, a 1 or 2TB HDD with such a cache (32/64GB) would be excellent, however, no manufacturer has offered such product, limiting themselves to 4/8GB flash, which is insufficient.
With no major OEM support that was never more than an enthusiast gimmick; and even at the time capping it at 64gb was a major limit. At that time the i7-2600 launched I already owned a 120GB drive; not being able to use all of my expensive purchase meant that even if I'd upgraded to a newer CPU I wouldn't've taken advantage of it. (And minimizing the hassle of multiple drives meant I bought a $370 240GB SSD about 18 mo after paying $300 for 128GB; and spent $500 on a 1 TB one two years ago so I could go pure flash in my desktop.) If I could've hybridized a larger SSD with an HDD into a single volume I'd've probably jumped on the opportunity to do so.
It actually has nothing to do with performance as for most users (people who don't read this site). Any of the most recent 8 GB NAND based SSHDs are actually good enough for most people's usage. What actually killed it off was the price to OEMs, which is really where client storage thrives. OEMs do not want to spend that extra money on the SSHD, especially when they see the average joe buying either their lowest end products which are cheap and have HDDs, or the higher end products including macbooks (SSD only now) and "Ultrabooks" and gaming PCs (pretty much becoming SSD only) where they have the budget for an SSD. The SSHD market hits a very very small market, which is even a niche in itself within the enthusiast group, since most of us would just have an SSD OS drive and HDD secondary drive. And the client PC market has been shrinking for a while now, so an already tiny market becomes even smaller.
the problem for seagate is not entirely SSD, Seagate didnt' R&D and up capacity of their HDD lineup enough.. They added 2TB every 4 years. If they can get capacity up to like 16TB-32TB per drive with new materials, and technology, then HDD's will make a return to the market.
I don't think we'll ever see 16-32TB in a spinning disk. In a SSD - definitely we'll see it. They don't even fill a 2.5" frame completely now, so the extra space can be filled with more multi-layer memory chips to create some multi-TB behemoths. Hey, maybe they'll even create a 3.5" body for fitting in more chips on the board.
Of course, they'll also need a controller to handle all of that addressable space, but I'm sure that's a hurdle that will be overcome.
Once you increase the density of HDDs with new tech is pretty easy to hit tens of TB's on them. From non PMR to PMR drive the density increase by near 10x. HAMR is/was that next stepping stone. Then we would have next gen HAMR for the 1st petabyte class drive.
SSD's will be expensive as multiterabyte storage for a long time.
The thing is, hardly anyone needs that much space. Hard drive capacity has far outstripped the demands of most users, and SSDs are quickly catching up. Just having massive drives may help them in the data center, but it won't make that much difference among end users. For the vast majority of people a 3 TB drive or a 300 TB drive are equally excessive.
You can get data-center pull Hitachi 2TB drives that are rock-solid for 50 USD or less (usually less).
The problem with the lack of price lowering is that (well made) hard-drives, unlike SSDs, don't actually die linearly (SSDs have a very finite lifespan for reliable use), meaning that the massive amount of hard drives that were deployed at data-centers flood the market after the ~3 year replacement cycle.
This makes it very hard to keep selling more hard drives as well made/designed hard drives can last up to 10+ years fairly consistently.
No bi deal they can join the 94 million not working in this wonderful economy, in no time their jobs will be filled with people from india and mexico at 1/3 the wages.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
41 Comments
Back to Article
plopke - Thursday, July 14, 2016 - link
That cost did not go up much compared to amount of extra fired people, was kinda on the horizon that this would happen, good luck to them.Did WD start to cut already, I mean they have SANDisk, HGST,.... so many things still merging?
Lolimaster - Thursday, July 14, 2016 - link
Capitalism as its best."Small" competitos hit your agenda? Buy them.
Remember:
"Mundo never merge! Mundo acquire!"
Chaser - Sunday, July 17, 2016 - link
Do share with us how other social systems would have increased hard drive demand.ddriver - Thursday, July 14, 2016 - link
So apparently, even cutting warranty down to the laughable 12 months their business model still sucks, thus the need to kick out people who have nothing to do with it, while the lousy management lingers on...Murloc - Thursday, July 14, 2016 - link
SSDs killed HDDs, other companies took a big part of the memory business away from the incumbents, and you blame management?DanNeely - Thursday, July 14, 2016 - link
For being too slow to realize that SSDs were going to disrupt the HDD market and not moving into the new business fast enough, yes I do blame Seagate's management. Along with WD they're hemorrhaging HDDs sales and not bringing in nearly enough SSD to replace them. I think Toshiba is better off here; but that has as much to do with being a bit player in the HDD market as to being a major player in the flash one.romrunning - Thursday, July 14, 2016 - link
Completely agree here. They simply didn't invest enough in SSD. For that, you can definitely blame mgmt. Five years ago, they should have seen what was coming, and started their investing then. They could have been co-leaders in the retail/consumer SSD market along with Samsung.Of those 8,000 jobs being cut, how many are the higher-up ones who lacked the foresight & also didn't want to change direction when the tide was clearly turning?
romrunning - Thursday, July 14, 2016 - link
I say five years ago, but I guess they should have started investing serious money as soon as they saw the mainstream Intel X25-M doing well in 2008. That alone should have told them where the future was going.Oxford Guy - Friday, July 15, 2016 - link
Sandforce, though, provided a moderate performance platform so the only thing left over was the really low end and the higher end. Even Intel ended up producing Sandforce drives. I'm not sure what Seagate would have been able to do, other than try to beat Samsung to the market. The 830 series were good drives. And the lackluster 840 TLC stuff was massively hyped and was a major market success.I suppose Seagate could have bought Sandforce instead of having them be closely tied to OCZ (the result of which was highly unstable drives like the 64-bit NAND Vertex 2 240 sold with false specs on the packaging) — and a horrible "panic mode" designed to destroy the ability of a user to get their data so that Sandforce would sleep peacefully knowing no one would be able to pry into its subpar firmware that doesn't work with TRIM correctly.
romrunning - Friday, July 15, 2016 - link
Well, as a counterpoint, the current market leader - Samsung - uses their own controllers. They didn't resort to Sandforce. Even if Seagate didn't want to partner with SandForce, they could have partnered with Marvell to use their controllers while they were building their own internal expertise in controllers. Instead, they did neither and mainly tried to stay status quo with HDDs.Meanwhile, SSD prices were going down and performance/capacity was going up. Seagate stays head-in-sand and still doesn't try to get a complete SSD product line. Samsung fills out their product line with TLC-based drives along with already well-performing MLC drives, and pretty much becomes the market leader. Even WD buys into their own future by acquiring Sandisk, thus ensuring they'll still have life after HDDs really tank. I think Seagate should be looking at acquiring Kingston or Micron to keep their future vibrant rather than just a slow decline.
Ariknowsbest - Monday, July 18, 2016 - link
Seagate bought SandForce from Avago in 2014, so they currently have in house controller just as Samsung.The only viable option for Seagate is to aquire Micron or SK Hynix. Buying Kingston wouldn't add any value to Seagate, as they can just as well buy the nand from the market and package it.
DanNeely - Monday, July 18, 2016 - link
Going after Micron as a whole probably wouldn't make sense; too big (about 50% higher market cap than Seagate ~13bn vs 9bn) and has too many other business units that wouldn't be of any interest (eg dram). Since the flash production is via a joint subsidy they co-own with Intel it's possible the latter might have a poison pill in place to protect its share of the investment.SK Hynix is even bigger (~ 22bn USD). Being part of the 3rd largest conglomerate in South Korea the barriers to any sort of takeover are probably insurmountable. (South Korean holding structures tend to look like Rube Goldberg machines and are designed to let the founders have majority voting rights even though they own a much smaller fraction of the total shares on the market.)
Lolimaster - Thursday, July 14, 2016 - link
And HAMR is nowhere to be seen. Was supposed be the zomg of drives 2 years ago...Lolimaster - Thursday, July 14, 2016 - link
15-20-30TB drives, hello?Lolimaster - Thursday, July 14, 2016 - link
They had the hybrid drives. 5 years later and they're still stuck with a laughable 8GB of really bad performing nand.Samus - Friday, July 15, 2016 - link
Yep, that was a great concept to cache high io hits, independent of the file system. But for some reason they never really perfected it with more NAND and enabling write caching.Firebat5 - Saturday, July 16, 2016 - link
I have to say I decided against the hybrids because I didn't think their NAND caches were big enough.stargazera5 - Thursday, July 14, 2016 - link
SSDs are a major part of it, no doubt, but you should also look at what gave SSDs such a large opening to start killing off their business so soon. I'm talking about the price-fixing in the wake of the 2011 Thailand floods. The HDD manufacturers enjoyed a turn around from the price war they had been fighting and artificially kept the prices elevated when they should have dropped after the flood. IMHO today they still haven't returned to the rapid price drops they had before. People being pushed to pay more changed 3 behaviors:1. People moved to cloud storage to offload the risk of HDD prices, and cloud providers could consolidate the number of disks used in a way many individual users couldn't.
2. With the price premium for SSDs dropped, people who could afford it started moving to them for their main drives instead of HDDs
3. End users kept disks in use longer then they would have before and brought older disks that had been retired back into service to fill the gaps.
All this reduced the market for HDDs and the manufacturers are now paying the price. Add in the collapse of the PC market as equipment cycles were extended across the board, and you can see why they are in trouble.
hechacker1 - Saturday, July 16, 2016 - link
In addition to cloud storage, cloud providers (streaming and document storage, Office 365, etc) killed the need to store a lot data locally. Enterprises are moving a lot of infrastructure to the cloud.Lolimaster - Thursday, July 14, 2016 - link
For avrg joe and jenny who barely store a decente amount of big a*ss mkv's. YES. SSD killed them.240-480GB and they're happy.
Samus - Friday, July 15, 2016 - link
How will Joe and Jenny's edit their amateur videos with less than 240GB!?Samus - Friday, July 15, 2016 - link
Seagate is the Chrysler of hard drives, without the long warranty. Would you buy a Chrysler with a 12 month warranty?HomeworldFound - Thursday, July 14, 2016 - link
Are they even making money on consumer products?DanNeely - Thursday, July 14, 2016 - link
According to articles Anton's written about the HDD market, low end consumer drives do return a very small return above manufacturing costs.But it's the expensive max capacity and enterprise models that generate the large marginal profits needed to fund R&D; which's why the slump in $50 laptop sales looks a lot worse from sales volume numbers than it is from a profitability/R&D impact perspective. OTOH when SSDs manage to gut the 10/15k SAS market; that will hurt their ability to sustain R&D a lot more. (And with Seagate thinking its next 15k generation might be the last at that speed the beginning of the end might be near.)
JoeyJoJo123 - Thursday, July 14, 2016 - link
IT'S OGRE. HDDS ARE DEAD. SSDS WON.JUST GIVE IT UP.
Sushisamurai - Thursday, July 14, 2016 - link
I'm surprised there's not a lot of coverage to those SSD/HDD hybrids. Did they not catch on and die off or have they not improved significantly?HomeworldFound - Thursday, July 14, 2016 - link
No real improvement unfortunately. There's just not a scenario where it makes sense. If the SSD side were bigger and you could choose what you wanted to store on there, it might work.DanNeely - Thursday, July 14, 2016 - link
They've more or less withered and died. The original models were things like 4/256GB. The 4GB of flash was too small to be useful for much beyond a write cache; and even there the simplistic write tens of gigabytes of data testing that was common at the time meant that it wasn't benchmark visible. At the time a 32/64GB flash 500gb laptop HDD (or 2/4TB desktop HDD) combo might've offered enough flash to hit the sweet spot if coupled with software/drivers smart enough to cache OS/Application binaries and frequently accessed user data on flash while archiving apps that haven't been used for months and media to the HDD part. Only Apple actually delivered a hybrid drive with that sort of resource mix; never having used one I can't comment on how well their software delivered.Now the problem for relevancy is that an SSD big enough to be suitable for most users is about the same cost or cheaper than a smaller but still useful SSD and a small laptop HDD. The only groups who still might need something like that would be gamers on a budget and people with large media collections they don't want to offload to a NAS. The problems there are both that those are relatively small groups of users, and that they're also the users least likely to find having to manage multiple drives a serious burden.
frenchy_2001 - Thursday, July 14, 2016 - link
Intel has offered ways to pair SSD/HDD in that way, with up to 64GB of SSD caching most used data automatically in your HDD. I've been using this solution for years (this came our of the z68 chipset, core i 2xxx generation). This works ridiculously well with only 32GB for the cache.So, a 1 or 2TB HDD with such a cache (32/64GB) would be excellent, however, no manufacturer has offered such product, limiting themselves to 4/8GB flash, which is insufficient.
DanNeely - Thursday, July 14, 2016 - link
With no major OEM support that was never more than an enthusiast gimmick; and even at the time capping it at 64gb was a major limit. At that time the i7-2600 launched I already owned a 120GB drive; not being able to use all of my expensive purchase meant that even if I'd upgraded to a newer CPU I wouldn't've taken advantage of it. (And minimizing the hassle of multiple drives meant I bought a $370 240GB SSD about 18 mo after paying $300 for 128GB; and spent $500 on a 1 TB one two years ago so I could go pure flash in my desktop.) If I could've hybridized a larger SSD with an HDD into a single volume I'd've probably jumped on the opportunity to do so.metayoshi - Thursday, July 14, 2016 - link
It actually has nothing to do with performance as for most users (people who don't read this site). Any of the most recent 8 GB NAND based SSHDs are actually good enough for most people's usage. What actually killed it off was the price to OEMs, which is really where client storage thrives. OEMs do not want to spend that extra money on the SSHD, especially when they see the average joe buying either their lowest end products which are cheap and have HDDs, or the higher end products including macbooks (SSD only now) and "Ultrabooks" and gaming PCs (pretty much becoming SSD only) where they have the budget for an SSD. The SSHD market hits a very very small market, which is even a niche in itself within the enthusiast group, since most of us would just have an SSD OS drive and HDD secondary drive. And the client PC market has been shrinking for a while now, so an already tiny market becomes even smaller.Morawka - Thursday, July 14, 2016 - link
the problem for seagate is not entirely SSD, Seagate didnt' R&D and up capacity of their HDD lineup enough.. They added 2TB every 4 years. If they can get capacity up to like 16TB-32TB per drive with new materials, and technology, then HDD's will make a return to the market.romrunning - Thursday, July 14, 2016 - link
I don't think we'll ever see 16-32TB in a spinning disk. In a SSD - definitely we'll see it. They don't even fill a 2.5" frame completely now, so the extra space can be filled with more multi-layer memory chips to create some multi-TB behemoths. Hey, maybe they'll even create a 3.5" body for fitting in more chips on the board.Of course, they'll also need a controller to handle all of that addressable space, but I'm sure that's a hurdle that will be overcome.
Lolimaster - Thursday, July 14, 2016 - link
1st gen HAMR was supposed to bring 15-30TB by 2018 (if released in 2015). And options up to 100TB by 2025.Lolimaster - Thursday, July 14, 2016 - link
Once you increase the density of HDDs with new tech is pretty easy to hit tens of TB's on them. From non PMR to PMR drive the density increase by near 10x. HAMR is/was that next stepping stone. Then we would have next gen HAMR for the 1st petabyte class drive.SSD's will be expensive as multiterabyte storage for a long time.
Nagorak - Tuesday, July 19, 2016 - link
The thing is, hardly anyone needs that much space. Hard drive capacity has far outstripped the demands of most users, and SSDs are quickly catching up. Just having massive drives may help them in the data center, but it won't make that much difference among end users. For the vast majority of people a 3 TB drive or a 300 TB drive are equally excessive.Lolimaster - Thursday, July 14, 2016 - link
Before the flood gate you could buy a 2TB drive for like 70 bucks. Now it's even more expensive.4 WD Blue 6TB for my personal data center.
Communism - Friday, July 15, 2016 - link
You can get data-center pull Hitachi 2TB drives that are rock-solid for 50 USD or less (usually less).The problem with the lack of price lowering is that (well made) hard-drives, unlike SSDs, don't actually die linearly (SSDs have a very finite lifespan for reliable use), meaning that the massive amount of hard drives that were deployed at data-centers flood the market after the ~3 year replacement cycle.
This makes it very hard to keep selling more hard drives as well made/designed hard drives can last up to 10+ years fairly consistently.
Communism - Friday, July 15, 2016 - link
Here is one for 30 USDhttp://www.ebay.com/itm/Hitachi-Ultrastar-7K3000-H...
And with 14,736 sold from just this one listing, you can see the scale of this problem for manufacturers.
svan1971 - Friday, July 15, 2016 - link
No bi deal they can join the 94 million not working in this wonderful economy, in no time their jobs will be filled with people from india and mexico at 1/3 the wages.bludothesmelly - Thursday, July 21, 2016 - link
Not a fan of seagate. Hard drive didn't last over a year.