News & Analysis
Comment
dougcp444
I think the ones with IP, like STEC, Sandforce, Sandisk, will survive and/or be ...
dougcp444
It's already been happening, I think. You can't find a major, and maybe minor, ...
SSDs: Still not a 'solid state' business
Mark LaPedus
8/20/2010 2:23 PM EDT
SANTA CLARA, Calif.—At one time, the next big thing that was supposed to drive the NAND flash market was solid-state storage drives (SSDs).
Beyond being the next big driver for the NAND market, some at one time also went as far as to say that SSDs would replace traditional hard disk drives (HDDs) in PCs. Based on standard NAND flash, SSDs are low-power, storage devices that could threaten power-hungry HDDs.
While SSDs have seen steady growth over the years, they have to some degree been a disappointment, with the industry still mired in the early adoption phase. SSDs are still somewhat expensive for cost-sensitive consumers and there is a perception that these products have some inherit reliability issues in the enterprise.
Thanks to the recent downturn, NAND vendors were reluctant to drive down their chip prices as fast as before, putting SSDs even further behind the cost delta curve behind traditional HDDs. Prices for NAND chips have suddenly fallen, but it could be too little and too late for SSD vendors.
This is not to say the SSD market is a total bust. But now, the SSD crowd has given up the notion that SSDs will totally replace hard drives. Some wonder if SSDs will even find any mass appeal at all.
SSDs ''won't replace hard drives,'' said Walter Fry, distinguished member of the technical staff at Hewlett-Packard Co., during a panel at the Flash Memory Summit here this week. ''Hard disk drives will continue to lead in cost per capacity.''
Others say SSDs will have a more important role outside of the PC. During a keynote address, Andy Walls, technical lead for IBM Systems and Technology division's deployment of SSDs, argued that SSDs will play an important role in data centers.
Jim Handy, an analyst with Objective Analysis, said SSDs are growing at about the rate he expected, but that this rate is slower than many had hoped.
"People had false expectations [for SSDs]," Handy said. "The PC market for SSDs has been slow to develop. The strongest growth has occurred in areas where HDDs simply will not operate and in systems for which users are willing to pay a significant premium for an SSDs' faster speed or greater durability."
Still, the market is growing in select segments. Some 4 million SSDs will ship in 2010, according to Objective Analysis. The firm predicts that nearly 40 million SSDs will ship in 2015, accounting for more than $7 billion in revenues.
Back in 2007, research house International Data Corp. (IDC) forecast that the SSD market would hit $5.4 billion by 2011 and that mass adoption would occur by 2011 or 2012. But this week, Mario Morales, an analyst with IDC, admitted that mass market adoption of SSDs is still "a generation or two away."
And SSDs have not exactly driven NAND demand, as previously predicted. SSDs are expected to make up only 6.1 percent of worldwide NAND bit shipments in 2010, according to IDC. By 2013, SSDs are expected to make up only 9 percent of worldwide NAND bit shipments, according to the firm.
That pales in comparison to mobile phones, which are expected to make up 36.1 percent of NAND bit shipments this year and 38.9 percent in 2013, according to IDC. In 2010, mobile phones are expected to be the biggest market for NAND bit shipments, followed by MP3 players (21.5 percent), digital cameras (17.3 percent), USB drives (10.5 percent), and then SSDs.
Next: SSDs still hold promise
Beyond being the next big driver for the NAND market, some at one time also went as far as to say that SSDs would replace traditional hard disk drives (HDDs) in PCs. Based on standard NAND flash, SSDs are low-power, storage devices that could threaten power-hungry HDDs.
While SSDs have seen steady growth over the years, they have to some degree been a disappointment, with the industry still mired in the early adoption phase. SSDs are still somewhat expensive for cost-sensitive consumers and there is a perception that these products have some inherit reliability issues in the enterprise.
Thanks to the recent downturn, NAND vendors were reluctant to drive down their chip prices as fast as before, putting SSDs even further behind the cost delta curve behind traditional HDDs. Prices for NAND chips have suddenly fallen, but it could be too little and too late for SSD vendors.
This is not to say the SSD market is a total bust. But now, the SSD crowd has given up the notion that SSDs will totally replace hard drives. Some wonder if SSDs will even find any mass appeal at all.
SSDs ''won't replace hard drives,'' said Walter Fry, distinguished member of the technical staff at Hewlett-Packard Co., during a panel at the Flash Memory Summit here this week. ''Hard disk drives will continue to lead in cost per capacity.''
Others say SSDs will have a more important role outside of the PC. During a keynote address, Andy Walls, technical lead for IBM Systems and Technology division's deployment of SSDs, argued that SSDs will play an important role in data centers.
Jim Handy, an analyst with Objective Analysis, said SSDs are growing at about the rate he expected, but that this rate is slower than many had hoped.
"People had false expectations [for SSDs]," Handy said. "The PC market for SSDs has been slow to develop. The strongest growth has occurred in areas where HDDs simply will not operate and in systems for which users are willing to pay a significant premium for an SSDs' faster speed or greater durability."
Still, the market is growing in select segments. Some 4 million SSDs will ship in 2010, according to Objective Analysis. The firm predicts that nearly 40 million SSDs will ship in 2015, accounting for more than $7 billion in revenues.
Back in 2007, research house International Data Corp. (IDC) forecast that the SSD market would hit $5.4 billion by 2011 and that mass adoption would occur by 2011 or 2012. But this week, Mario Morales, an analyst with IDC, admitted that mass market adoption of SSDs is still "a generation or two away."
And SSDs have not exactly driven NAND demand, as previously predicted. SSDs are expected to make up only 6.1 percent of worldwide NAND bit shipments in 2010, according to IDC. By 2013, SSDs are expected to make up only 9 percent of worldwide NAND bit shipments, according to the firm.
That pales in comparison to mobile phones, which are expected to make up 36.1 percent of NAND bit shipments this year and 38.9 percent in 2013, according to IDC. In 2010, mobile phones are expected to be the biggest market for NAND bit shipments, followed by MP3 players (21.5 percent), digital cameras (17.3 percent), USB drives (10.5 percent), and then SSDs.
Next: SSDs still hold promise
|
|
|
|
|
Navigate to related information
unknown multiplier
8/20/2010 10:12 PM EDT
An HDD is heavier, consumes much more power, not to mention slower than SDD. Even if SSD fails to break into the server sector or the notebook sector, it could still have an opportunity for devices like netbooks and smaller/lighter. But the competition there would be cards instead of HDDs. But at least it still uses the same underlying NAND Flash memory technology.
Another thing you have to consider is the product replacement cycle. If it is something like 3 years, GB/TB storage may not be practical since you may not store so much in your device in just three years.
Sign in to Reply
resistion
8/21/2010 12:02 PM EDT
SSD is limited to computers (including notebooks) only not smartphones. Smartphones are themselves like USB flash drives or use SD cards. They wouldnt have HDDs in them to replace in the first place. I think even ipads don't have SSDs, just flash drives.
Sign in to Reply
Dave.Dykstra
8/21/2010 1:39 AM EDT
Despite the numerous advantages of SSDs, they will not become mainstream until either the cost differential is relatively insignificant (because users are typically not willing to pay a "significant premium" for SSDs), or the devices being used change significantly enough that those with SSDs are the mainstream devices. Some recently introduced devices using SSDs have sold very well, but the numbers involved have not yet reached the point where these devices are the mainstream devices. It appears that while pricing of SSDs will continue to come down, sales of devices using them is also continuing to increase and the combination of those two things should drive increased usage of SSDs at a somewhat more rapid rate than only one of those factors would indicate.
Sign in to Reply
mark.lapedus
8/21/2010 10:39 AM EDT
Thanks for the comments. Do we have more debate on SSDs vs. HDDs? I say SSDs won't become a flop as say optical. But SSDs will become a niche.
Sign in to Reply
KB3001
8/22/2010 12:31 PM EDT
I migrated to SSD last year and I have not looked back since. A SSD is much smaller than HDD and, more importantly to me, much faster. I did not mind the price tag to be honest for such features. As for the future, I think SSDs will find more use in mobile and handheld devices as these become more and more complex and much like a desktop PC in terms of their functionality. I am not sure a less than projected sales figures in the traditional PC market is a source of major concern. There are other issues too these days to do with flash production capacity. All in all, I reckon it's just a blip and the future is SSD for me.
Sign in to Reply
VincePG
8/22/2010 1:52 PM EDT
I think the author of the article is too hung up on cost. Today and for the foreseeable future SSD will lose to just Flash or HDD on cost, so today, if the SSD market is going to take off it needs a market that exploits its strengths. Tomorrow, since storage is not like performance in 3D graphics, at some point more storage reaches a point of diminishing returns. At that point SSD will over take HDD in a particular market. SSD(the distinction between just Flash and SSD being a standardized interface SATA, m-SATA, Fibre Channel, SAS, and ATA/IDE), separates itself from HDD under four important criterion, low power, fast access, size and durability in g-force environments. SSD separates itself from just Flash, through a scalable, standard interface. Low power and fast access have been widely discussed and are important, but I don’t believe are the driving factors and this fact is proven by the market numbers. If low power and fast access are your only concern just put in flash. That’s what mobile does today. The real promise for SSD today is expandable memory in harsh environments. Applications such as black boxes for impact, such as cars, boats and bicycles(they’re already used in airplanes), expandable DVR functions for mobile phones( Samsung introduced an HDTV mobile, DVR must be coming), and memory for video cameras are three examples. Worldwide emphasis on security against terrorism where blast and crush survivability of cameras storage devices is another huge growth arena for SSD. I wouldn’t call SSD a niche product. I’d call it an evolutionary product that will find more markets as HDD reaches its point of diminishing returns.
Sign in to Reply
Rick Merritt
8/22/2010 6:58 PM EDT
Neither HDDs or SSDs are likely to die or become niche in the next ten years. In fact, both are likely to continue to expand as digital technology expands with new products and markets and both technologies improve.
I wrote a couple stories this week about the long term road map for HDDs. See http://www.eetimes.com/electronics-news/4206326/Decision-time-looms-for-hard-drive-makers
The upshot is drives have a major tech transition ahead in about 2015 and much is still uncertain. Still HDD makers have been through this sort or re-invention before, there are plenty of ideas on the table and drives are likely to continue costing a few pennies per Gbyte into the foreseeable future.
Sign in to Reply
phoenixdave
8/22/2010 8:38 PM EDT
@Rick - I totally agree that both are likely to be around for a long time. They both have advantages for their applications. There's room for both in this fast-changing technology world.
Sign in to Reply
mark.lapedus
8/23/2010 3:52 AM EDT
I agree too. However, Gartner thinks that 90 percent of all SSD makers will fail. That makes sense. There are too many SSD makers out there. The big guys (Intel, Samsung, etc.) will survive. These guys have fabs and controller technology.
Some say the guys with no fabs or controller technology will die. Agree or not?
Sign in to Reply
dougcp444
10/18/2010 11:09 PM EDT
I think the ones with IP, like STEC, Sandforce, Sandisk, will survive and/or be bought
Sign in to Reply
andyzg
8/23/2010 4:41 AM EDT
Yes. SSDs will be integrated maybe onto a single SOIC with PCIE/AHCI or SATA. not necessarily large, it is enough to hold C: drive contents (windows and apps). if the user needs bulk storage, they may have a NAS or a mechanical HDD can be added as drive D:
Sign in to Reply
resistion
8/23/2010 5:46 AM EDT
Enterprise storage may be the next meaningful battlefield for HDD vs SSD. But it doesn't look like HDD will lose, more of a matter of how much SSD can win.
Sign in to Reply
mark.lapedus
8/23/2010 12:48 PM EDT
Right. Agree. But are the Fortune 500s buying SSDs in their datacenters? Can they trust them? If I ran a poll today, I'll bet they would say no.
Sign in to Reply
dougcp444
10/18/2010 11:06 PM EDT
It's already been happening, I think. You can't find a major, and maybe minor, OEM (HP, IBM, EMC, Oracle, Dell, and many smaller OEMs) that do not have SSDs in their large product lines. I think if you took a tour of these companies product announcements of this year you might be pretty surprised. I would expect to see SSDs in many cloud architectures, too.
Sign in to Reply
dirk.bruere
8/23/2010 3:52 PM EDT
SSDs will come to share desktop PCs with HDDs.
My next PC will include a 128G SSD for OS and key apps, and a few TB of storage for media etc. However, the place where SSDs will make a big impact is laptops. My next laptop will be SSD only - I don't need TB of data on it. Also, my company is getting rid of the HDD from its media products and replacing it with SSD and HDD NASboxes
Sign in to Reply
JamesAndersonMerritt
8/23/2010 4:02 PM EDT
Price is a big issue, in terms of whether or not there is a consumer business or market there. For myself, if I could get a 200GB SSD for $300-400, I'd buy it immediately, as replacement of the cranky HD now in my laptop. But such drives now retail for hundreds of dollars more. I'm biding my time, and I bet that many others are, too.
Sign in to Reply
przemek
8/23/2010 4:20 PM EDT
Besides high performance, another big advantage of SSDs in a high-end datacenter environment is their well-characterized life cycle. Rotating media hard drives fail randomly, whereas SSDs have a predictable endurance profile The firmware can estimate how much of the write endurance budget remains. This enables reliable predictive maintenance schedule and cuts down weekend overtime for the harrowed system administrators.
Sign in to Reply
LarryM99
8/23/2010 4:31 PM EDT
I'm not used to being in the majority, but it looks like a lot of people here agree with me. If you track what is going on in the gaming PC space you see that the case manufacturers are starting to add slots for 2.5" devices. Their customers are demanding SSD's because they have figured out that they provide a major performance boost for disk I/O, especially on boot times.
Does that translate to use by sane people? :-) At current prices, maybe not. But I have been running one in my main machine for a few months now. I am getting used to clicking on Outlook and having it come up practically instantly. I am spec'ing out a new HTPC right now and planning on an SSD boot drive (quiet, fast, and low power). I will be complementing it with a 2TB hard drive, though. I may be crazy but I'm not stupid.
The error is in thinking of it as 'either-or'. SSD's will replace mechanicals in some applications, but will generally complement them. The real gains will show up when they stop pretending to be traditional hard drives, though, and are directly supported by the OS.
Larry M.
Sign in to Reply
docdivakar
8/23/2010 4:31 PM EDT
@Mark, you have a dittohead here :-) on your statements that SSD makers without the fabs & controller technology will die, and major datacenters are not jumping to adopt SSD's yet. I agree on these and many other comments posted here.
In the days leading up to virtualization and also in highspeed cluster computing, the interconnect pipelines (Cat6 copper) got a bad rap for being too slow and adding to the latency. But now every one knows that even in a gigabit cluster/virtualization environment, the interconnects are only adding to about 10% of the end-to-end latency; the rest comes from (among others) accessing data including reading drives, applications, and cluster / virtualization operating systems (like MPI's & hypervisors).
SSD's therefore can add a significant advantage in latency-sensitive applications (like in financial market, transaction processing, etc.). But their reliability has to be proven beyond lingering doubts.
Sign in to Reply
jimcondon
8/23/2010 7:59 PM EDT
I've seen several papers pushing the power usage advantage of SSD's in data centers. I think as the push comes to support more and more green data centers, you'll see a substantial growth in the use of SSDs there.
In the long run however I don't think you'll ever get rid of HDD. For the foreseeable future, HDD will outpace SSD on price per GB and that in itself will win in some markets. Very similar to how the desktop won't ever be replaced totally by laptops. Although the laptop is smaller and more convenient, the cost advantage of the desktop will win some markets.
Sign in to Reply
goafrit
8/23/2010 9:05 PM EDT
Good point. Your analog with laptop and desktop explains this very well. I think so. When two years ago the SSD was on, I knew it was just to have a balance and never to displace HDD.
Sign in to Reply
mark.lapedus
8/23/2010 9:27 PM EDT
But here's the question. Will datacenter managers implement SSDs in their facilities? Some will. I say the vast majority won't. Who will risk their jobs on it?
Sign in to Reply
CamilleK
8/23/2010 10:18 PM EDT
I think SSD will be bigger than currently projected. The issue on higher cost will be made insignificant after a certain adoption rate is reached, in that price drives volume (negatively) until volume drives price (positively) and no one will look back. Just on the technical merits(speed, power, form factor size, no mechanical components, noise, green factor)the edge will start going the SSD way.. Even if endurance is still to be conquered, the replacement cycle will be less onerous for large data centers. When HDDs will have the same footprint, are faster and consume less power, I may change my mind. SSDs still need some software enhancements and there is already too much legacy working against SSD. It is not easy to incrementally add SSD to HDD from a maintenance standpoint, but new data-center build out modules could be pure SSDs.
Watching those 1500 video HD streams coming out from one SSD module (demo'ed on the flash summit exhibit floor by Fusion IO) got me jazzed up.
Sign in to Reply
Peter Liu
8/23/2010 10:44 PM EDT
For SSD adoption as mainstream, one question people forget to ask is: how long will you use your PC before you buy a new one and during the life cycle of your PC, how much storage do you really need? One average, the PC life cycle is 3-5 years and 128GB storage is more than enough. So if 128-200GB SSD has the same price as 1TB or even 2TB HDD, the extra HDD space is useless for most people.
I attended this Flash Memory Summit, what surprised me is none talked about the impact of cloud storage on the adoption of SSD as the mainstream. My opinion is: with the population of cloud storage and higher network bandwidth, we will need less storage space on the PC, and in 5 years, most of us will not see HDD on laptop or desktop.
Sign in to Reply
unknown multiplier
8/24/2010 11:31 AM EDT
Yes, I agree totally. In 5 years, I think all of my internet downloads and media files will not be on a desktop or notebook.
Sign in to Reply
PierreMars
8/24/2010 12:19 AM EDT
My biggest fear when travelling is my HDD will crash if turbulence happens while working in a plane. My next notebook will have an SSD for that reason alone. I am sure there are many business travellers who will pay the premium for that robustness.
Sign in to Reply
GREAT-Terry
8/24/2010 12:38 AM EDT
I believe HDD won't die while it will be common to see more SSD in laptop but SDD (holding OS and demanding applications) complementing HDD in desktop. I've used SSD in my laptop and it is really very good to see its speed and roubustness. The only thing I'm concerning is how people are preparing to improve the number of read/write cycle in Flash. I think it is why datacenter peopler are so hesitating in adopting SSD. Once the robustness for holding data for decades is proven in Flash technology, the data center people surely will love to use it.
Sign in to Reply
mark.lapedus
8/24/2010 2:56 AM EDT
I agree with all points. But has anyone heard this yet: I need to run out and buy an SSD?
It's not on the top of my list. Thoughts?
Sign in to Reply
celectronic
8/24/2010 5:35 AM EDT
Not sure if this is the right question.
You may have heard the following:
Why does it need so much time to boot?
I need to get a laptop with longer runtime from the battery!
If I want to buy a new portable, battery runtime and boot speed are worth some money for me.
Sign in to Reply
sjprg
8/24/2010 9:26 AM EDT
At present I have two Intel X25-V in raid 0 in my desktop and a Vertex turbo in my laptop. I see SATA as a dead end for SSD with the future being PCI-E as the interface. Examining the Vertex Z series drives this is what I envision the future of SSD in the marketplace. With the advent of PCI-E 3 The HDD bottle neck may finaly be broken and external terrabyte will be USB3.
Sign in to Reply
MeirG
8/25/2010 5:58 AM EDT
There are two markets that I didn't see addressed:
1) The Home Entertainment Center PC: There, audible noise is my main concern. I don't want that when I listen to Elly Ameling singing Gretchen am spinnrade. And here, the competition is the cost of a "silent" PC enclosure. Have you seen the prices of these? And the home entertainment is not a niche market...
2) As a cache: IMHO, there is a slot between Main DRAM and HDD. It will than join the long chain from level-1 through 3 SRAM cache, SSD as the virtual memory and up to a robotic tape machine at the extreme off-line edge.
What say you folks?
Sign in to Reply
jimcondon
8/25/2010 8:28 AM EDT
MeirG, Great point on the Home Entertainment Center PC. Noise is another issue that people overlook as an advantage. I remember the first time I walked in one of my customer's offices that was outfitted with Silent PCs (no fans, SSD). I was stunned with how quiet it was and how much noise we take for granted out of our computers.
Sign in to Reply
mark.lapedus
8/25/2010 8:09 PM EDT
We all agree SSDs have a place. Should they be marketed better by the OEMS? If so, how would you market an SSD to expand the sales for these items?
Sign in to Reply
resistion
8/26/2010 2:54 AM EDT
The trouble is the SSD ATA interface only allows them to be marketed as HDD replacement. So you have to wait for everyone to be so unsatisfied with HDDs even at their lowest cost/bit that they will throw them out for something for expensiver, but lighter, faster, power-saving, etc. This takes too much time. SSD makers need to switch to an interface that allows HDD users to keep using HDDs while investigating the possibility of solid-state storage independently. The USB interface might be a good place to start. Capacities of 64 GB already available.
Sign in to Reply
mark.lapedus
8/26/2010 1:54 AM EDT
I received an interesting e-mail on SSDs. This commentary comes from Walker Blount, an analyst with Web-Feet Research:
''SSDs and flash storage appliances improving enterprise performance is dependent on the applications that are run and the architecture they are being deployed in. Usage patterns also factor into the mix where read intensive, write intensive, IOPs and bandwidth may be requirements of the application. Enterprise applications can be reduced to four fundamental areas: E-Business, Financial, Web searches, and Video. In each of these areas, understanding the application’s usage patterns becomes an important factor when deploying SSDs. In E-business and financial applications; On-Line Transactions (OLT), Analysis and Data Mining requires fast response times requiring low latencies. SSDs have this attribute whereas enterprise hard disk drives do not. Video applications such as Video on Demand, Webcasting, and Medical imaging requires high bandwidths to move large amounts of sequential data for viewing, which again are SSD attributes.
''SSDs and flash storage appliances in the enterprise are being deployed where enterprise hard drives dominate. They will either replace some or augment enterprise hard disk drives in the tiered storage environment. Determining where they will be deployed will depend on where the performance and cost benefit is best whether it is in application servers and/or network storage in the SAN environment.
''With the continued pressure for improved performance, reduced footprint, and reduced power consumption, IT managers must consider SSDs and flash storage appliances as performance and storage solutions. They will be challenged to find the best fit for these devices to provide the best performance and cost benefit for their enterprise applications and customers.''
Sign in to Reply
katgod
8/26/2010 2:29 PM EDT
I saw three SSD comments that need further exploration. One, crashing disk drive because of airline turbulence, are there known instances of this? Two, longer run time in portable computers, while this is a true statement what is the actual run time increase in average use. Three, are the new low noise HDDs really a problem for a home media center, I can just barely hear my drive when it is on my lap in my home, I am curious if a newer HDD in a box would be heard while listening to most source material. Yes, this might be a nice marketing feature for a high end home media center where cost would be less of an issue.
Sign in to Reply