Top SSD
Companies - © StorageSearch.com
based on reader search in
2nd Quarter 2012 |
| . |
|
|
| rank.... |
company |
notes........................................................... |
|
| 1 |
Fusion-io |
Same as before.
This is
Fusion-io's 14th straight quarter at #1.
Fusion-io's search volume
was nearly
2x the level of the #2 ranked company in this list and
5x the level of the #10 company and 10x the level of the #20
company. That seems like a big difference - but compared to previous editions of
this list - the gaps between the top 10 companies are shrinking.
2011
was the year of
the FIO IPO.
Now one of the big SSD trends in 2012 appears to be
the business impact of
DSP IP inside
flash SSDs.
So I recently asked Fusion-io to reveal what they
were doing with this new technology.
I got the answer that they didn't
want to disclose that level of detail right now.
OK - so why did I ask?
Isn't it obvious that at some stage in the future FIO (and nearly
everyone else in the flash SSD market) will use the new technology? It's just a
matter of timing... Some other enterprise SSD makers also declined to reveal
their exact plans in this area.
Yes - but let's re-examine why FIO is
in the market position it's in today. If you strip away its first mover
advantage in enterprise PCIe SSDs, its comfortable software fit with big server
architecture and its superb execution at marketing - Fusion-io is the company
which has the lowest component cost when it comes to delivering a PCIe SSD at
any given performance level.
Adaptive DSP IP enables SSD designers to
significantly lower the cost of their memory bill of materials and deliver
faster SSDs in the fast-enough
SSD segment - when using the cheapest flash.
Those factors
cancel out the cost, electrical power and reliability advantages which FIO's
(we don't need controller chips) minimalist hardware architecture has long had
over all its competitors in the fast-enough PCIe SSD segment.
The
result would be to level the competitive playing field in raw hardware costs
compared to companies like STEC.
There
are 3 ways FIO can deal with this challenge
- compete in the market without adaptive DSP IP. Customers who use the
company's software API's in
new dynasty
applications can achieve speedups which are 10x bigger than traditiona
API's. So why care about a mere 2x speedup from a new flash technology?
- FIO can develop its own adaptive DSP technology and integrate it within
its existing software stack. (My guess is this is what the company will do - or
may have already done.)
- FIO can license an IP set from another source - such as
DensBits. When
DensBits spoke to me last April I immediately saw that FIO would be the
broadest fit to adopt their technology from the technical point of view -
because both companies have very scalable architectures. The only downside for
FIO - if they choose this course - would be the licensing cost which is a
percentage of the whole SSD market price.
When and if FIO do publicly
announce their DSP IP adoption strategy - they'll go back to having the lowest
cost BOM - although if they're paying a license for the technology it may narrow
their competitive lead.
...Later:- 2 years later Fusion-io
was late to market with integrating smaller geometry flash memories within
its own IP set - which impacted its ability to compete.
See also:-
Why was
Fusion-io - the best known and most often admired enterprise SSD company -
unable to survive as an independent company? |
|
| 2 |
Violin Memory |
Same as before.
In this
quarter - Violin announced it had secured $50 million in Series D funding, which
the privately owned company said gave it at an extrapolated market value of more
than $800 million.
It's been a year since Violin established itself in
the minds of the market as the new trend setter in rackmount SSDs - a role
previously occupied by Texas
Memory Systems.
Like TMS - Violin uses its own
big SSD
controller architecture - although - unlike TMS - Violin doesn't offer any
SSD products in smaller form factors (yet) like PCIe SSDs.
In the past
year many other companies have entered the newly emerging market for
high availability SSDs.
And in the fast-enough enterprise SSD market - Violin's competition is also
getting tougher.
Can Violin sustain its lead in the rackmount market?
It will get more difficult as the rackmount SSD market itself
fragments more clearly into different types of products.
Violin has enough market momentum and presence to stay in the top
end of the SSD companies list - but I don't believe that any single company
can be a sustainable leader in more than 1 or 2 of the 5 headline segments
within rackmount SSDs. The technologies required and market experience are
too
diverse.
My guess is that from its vantage point Violin will cherry
pick the segments in which it does best.
Will adaptive DSP IP in SSDs
impact Violin?
Yes. And like FIO - Violin will be unable to use this
technology unless it develops the technology itself or licenses it.
I
think that before that happens - Violin's competitiveness (which come from the
efficiencies it has in
overprovisioning
ratios due to its big controller architecture) will be eroded by rackmount
competitors being able to use arrays of cheaper but relable DSP flash - in the
form of SAS SSDs. But
it could take another few quarters before Violin sees its cost advantages
shrink. |
|
| 3 |
STEC |
Up 1 place since the last quarter.
In
many ways this quarter has been a turning point for STEC.
In June I
said in our main SSD news
pages that the company had "improved its enterprise marketing" and
that "the company has doing new things which it didn't do before - rather
than just doing the same old things better."
That was after
years of eviscerating comments about the company's routes to market and
marketing.
So if you believe that marketing can have a beneficial
effect on company performance you should expect to see better results from the
company in the coming year (compared to what it would have achieved otherwise.)
In
this quarter - STEC announced the general availability of the company's
EnhanceIO SSD Cache Software for Linux and Windows environments.
I'm
not convinced about the long term sustainability of a hardware SSD maker
offering their software for use to accelerate competing SSDs. And I know that
many other SSD ASAP
products provide better performance - because they use more complex hot spot
algorithms. Nevertheless - STEC has opened the
SSD software gate as a
new portal for customers to do business with the company. And that will lead to
new business opportunities. |
|
| 4 |
OCZ |
Up 1 place since the last quarter. This
is OCZ's highest ranking in the history of these lists.
OCZ's SSD
revenue for its fiscal quarter which overlaps 2/3 of this calendar quarter
grew 54% year on year to reach $106 million. The immediate market
reaction was negative due to reported losses and a reader asked me to comment on
that while I was writing this piece.
I said my starting point for
evaluating the company would be this. If OCZ wanted to – it could sell its
controller business today -
Indilinx (which does
have its own adaptive flash DSP IP) for 20x as much as it paid for it.
So why worry?
OCZ competes across a wide spread of markets - but within
the enterprise market - it's tempting to make comparisons with STEC and
Fusion-io.
OCZ's SSD software set is much more capable than that from
STEC - but OCZ hasn't had more than a few quarters to integrate the powerful
software set it acquired from
SANRAD as a market
leveraging tool. That's in contrast to FIO for which software has been in its
DNA since founding. You can't use an ioDrive without the software. It won't
work. In contrast - you can use most of OCZ's SSDs without being tied to their
software.
Having said all that - due to the past history of OCZ, STEC
and Fusion-io each of these companies is nearly as different in its
accessibility to new customers as in its core technology - even when you're
looking at superficially
similar products like PCIe SSDs.
OCZ is the most accessible to
customers from a broad market. Fusion-io has had the most impressive set of
design wins and STEC - which had been for many years the most difficult company
to do business with - has recently changed the way it markets SSDs.
Each
of these companies have big strategic customers or markets which - for reasons
of business preference or technology risk aversion - they can serve better
than their "so-called" competitors. |
|
| 5 |
LSI/SandForce |
Down 3 places since the last quarter.
In
this quarter - LSI demonstrated its SandForce SF-2000 flash controllers
working with Toshiba 19nm and Intel 20nm NAND flash memory. And the company said
that its multi-generation roadmaps are a key reason that oems (currently 50
companies) like its way of doing things.
The 1st generation SandForce
controller supported 4 flash generations from 5Xnm down to 2Xnm and its current
generation started at 3Xnm geometries and already supports 3 generations of
flash down to 20nm.
LSI says that one of the reasons it can support so many generations of
flash from the same generation controller design is that it uses firmware as
part of the flash management IP.
However, a safe inference from the
market activities of some of LSI's oem customers is that in this quarter LSI
was not yet shipping adaptive DSP enhanced IP in its controllers.
...Later:-
It took another 18 months before LSI announced (in
November 2013) it
could deliver adaptive DSP in a new SSD controller - the SF-3700 and by the
time this product was sampling in 2014 - LSI had begun to lose considerable
share in the market for controllers aimed at enterprise 2.5" SSDs.
The
grab for adaptive R/W enterprise IP led to several
acquisitions:-
SMART Storage (by
SanDisk),
STEC (by
HGST / WDC) and
eventually LSI itself - whose SSD products were acquired by
Seagate. |
|
| 6 |
Texas Memory Systems... |
Same as before.
In this quarter - TMS didn't launch any new products - and instead
concentrated its marketing communications resources towards customer success
stories and collaboration with yet another SSD software company - DataCore.
Here's
a warning! - If you do a web search for the words "DSP" and "SSD"
you may get misdirected by mindless search-engines to some of TMS's earliest
projects with SSDs as data capture devives - or even to the company's recent
range of DSPs which are supported by SSDs.
Those juxtapositions of
the words "DSP" and "SSD" aren't what we're talking about
today. Although it is how some of us got into the SSD industry.
As I
was saying recently to a VP in a different SSD company who like me also had a
data acquisition background - we may be among the few people in the enterprise
SSD industry who understand what DSP means. (OK there are many thousands of
you in the industrial
and military markets
who know too.) |
|
| 7 |
SanDisk |
Up 1 place since the last quarter.
In
this quarter - SanDisk launched a new family of bootable enterprise PCIe SSDs -
the Lightning range - which leverages SSD IP from 2 previously acquired
companies (Pliant
for the controller hardware and
FlashSoft for the
auto caching software). |
|
| 8 |
Virident Systems |
Down 1 place since the last quarter.
Virident didn't launch any new products in this quarter.
One of the
contexts in which Virident has been mentioned in this quarter - is the validity
of different types of SSD benchmarks - and how some products have better
performance symmetry
than others.
The company has developed a new
SSD benchmarking tool
which would produce better
conformity between lab
benchmarks and the experience in real-world applications.
This is
one of many areas where the interests of better
SSD education
intersects with those of business development.
The general argument -
which most vendors would agree with - is that the more that users understand
SSDs - the more readily they will buy more SSDs.
But can you
extrapolate that to specific buyer behavior?
By that - I mean - if a
benchmark designed by a particular SSD company shows that their own product
works better than a competitor's product are you more likely to buy it? Even if
it's true?
Or will it simply mean that competitors will design their
own versions of benchmarks which make their own products look best? (It's easy
enough to pick preferential tests.)
In general I think that efforts to
increase the level of SSD education are a good thing athough the side effects
are:- it can give you a headache, or it can soak up your time, and you may be
even more confused at the end of the education process then at the start when
at least you were certain that you didn't know anything at all. |
|
| 9 |
WhipTail |
Up 2 places since the last quarter. This
is WhipTail's highest ranking in the history of these lists.
In this
quarter - WhipTail launched a multi-protocol HA SSD rack called the INVICTA
($250,000 for a 6TB 6U 250K IOPS 200 µS latency model).
WhipTail
is the highest ranking rackmount SSD company in the Top SSD Companies List -
which uses arrays of open
/ COTS SSDs instead of its own proprietary flash controller architecture. |
|
| 10 |
Kove |
Same as before.
In this quarter - Kove published new record latency numbers for its
fast RAM SSD - the XPD L2 - which achieved continuous and sustained 5
microsecond random storage read and write when connected via 40Gb/s InfiniBand
adapters from Mellanox . |
|
| 11 |
BiTMICRO |
Down 2 places since the last quarter.
In
this quarter - BiTMICRO announced it had obtained over 600 IP assets from
QualCore whose portfolio includes analog, digital, and mixed-signal IC design.
Engineers retained from QualCore's IP and ASIC services team have joined the
recently established BiTMICRO India. |
|
| 12 |
SMART |
Same as before.
In this quarter - SMART showed it can turn even more tricks by
leveraging its adaptive DSP IP inside SAS SSDs.
The 1st of these was
in preconditioning flash memory in unmodified
LSI/SandForce driven SSDs
(without adaptive DSP) - with new default write programming parameters
harvested from its DSP population - thereby yielding 5x better endurance.
The
2nd of these was another stretch of its DSP enhanced SSDs - in new models
which guarantee 50 full random drive writes per day for a period of 5 years.
(About 10x the previous industry benchmark established for this class of
SAS SSDs by STEC.) |
|
| 13 |
Anobit |
Up 1 place since the last quarter.
Although
Anobit was acquired by Apple last December that hasn't stopped people being
interested in what the company had been doing.
Why? You guessed it. It
was one of those pioneering DSP IP in SSD companies.
Gregory Wong founder of
Forward Insights
told me recently that his report -
ECC and Signal
Processing Technology for Solid State Drives and Multi-bit per cell NAND Flash
Memories 2nd Edition - which costs $6,500 - has been his best selling SSD
report. |
|
| 14 |
RunCore |
Up 2 places since the last quarter.
In
this quarter (among other things) - RunCore launched some pre-packaged MiniDOMs
leveraging its rSSD
(tiny SSD) aimed at the
industrial SSD
market.
And the company also released a seriously bad (but in an
interesting way)
video
- aimed at the consumer
SSD market - which redescribed its phone-zap and RFID
security technologies. |
|
| 15 |
Intel |
Up 2 places since the last quarter.
In
this quarter - Intel launched another new SSD aimed at the
consumer SSD market
- the
330
Series. |
|
| 16 |
DensBits |
First appearance in the top SSD companies
list.
When I spoke to Amir Tirosh
at DensBits in April I knew within a few minutes that DensBits would be a
company which would go straight from stealth mode into the Top SSD Companies
List - within the same quarter - and I stuck my neck out and said so to my
readers.
If you've been reading the earlier parts of this article -
instead of just scrolling down to spot who the new companies are - you've
already learned to appreciate the significant importance of that adaptive DSP
IP in SSDs thing.
DensBits is a company which offers SSD oems and
memory chip makers (but not controller companies) the combination to the magic
safe that holds these secrets.
Unlike some of the other SSD companies
who have also developed this technology for themselves for use in their own
specific markets - the DensBits tool box isn't limited in scale to vertical
markets. Their IP can be adapted (in different ways) to tune SSDs in consumer,
industrial and enterprise markets.
In this quarter -
Seagate announced it
will use DensBits's flash care technology in the design of forthcoming consumer
and enterprise SSDs. Seagate also made an equity investment in DensBits. |
|
| 17 |
EMC |
Down 2 places since the last quarter.
In
this quarter - EMC acquired
XtremIO for $430
million. |
|
| 18 |
WD |
Down 5 places since the last quarter.
In
this quarter - WD didn't launch any new SSD products or make any significant
SSD related announcements. |
|
| 19 |
Samsung |
Down 1 place since the last quarter.
In
this quarter - didn't launch any new SSD products or make any significant SSD
related
announcements. |
|
| 20 |
Nimbus Data Systems |
re-entry after an absence of 2 quarters.
In
this quarter - Nimbus didn't make any significant announcements. Its marketing
communications were all on the themes of performance validation, compatibility
etc.
In common with all
rackmount SSD
makers - which use SAS SSDs
inside their arrays - in the next few quarters we can expect to see speedup
stories from Nimbus and others which ride on the back of the 12Gbps broncos
which will probably be the final generation in this interface's roadmap.
(The
need for faster SAS
- beyond the 12Gbps generation will most likely be headed off at the pass by
2.5" PCIe SSDs
- although SAS will still be useful in less demanding parts of the enterprise
SSD silo.) |
|
| 21 |
Skyera |
First appearance in the top SSD companies
list.
For a company which was still in stealth mode during most
of this quarter - and which still hasn't yet launched any products - it was a
remarkable achievement to rocket straight in to the Top SSD Companies List -
even if it is towards the bottom end.
If you've ever read my March
2010 article - this
way to the petabyte SSD - you may have seen the press release from Exabyte
SSD Appliance which describes their 2U Petabyte SSD Appliance (2PB
uncompressed / 20PB usable). That article and press release describes the final
elements needed to complete the set of solid state storage parts in the
pure enterprise solid
state storage silo. And the article also describes the business case for
eliminating hard drives from the enterprise server farm - even if the price of
HDD capacity drops to zero.
To put that into context - the article was
published 5 months before Skyera was founded - and the company (Exabyte
SSD Appliance) - and the press release were spoofs which I set in the future.
But it was the simplest way for me to illustrate how the many complicated
strands of the SSD market would come together in a product concept to fit a
market niche which did not at that time exist - but would clearly materialize as
sure as night follows day.
Since that article was published it has
formed a useful part of the conceptual market backdrop in the many enterprise
SSD roadmap discussions I've had with leading industry thinkers.
Going
back to Skyera...
The simplest way to describe what I think Skyera is
doing is that they are designing bulk storage SSDs - which are on the roadmap
to delivering multi-petabyte scale SSD storage within 1 or 2 U rack space -
at a cost of ownership which will make using hard drives in the server farm
seem as old fashioned and quaint as using vaccuum tubes in a mainframe. That's
it. We'll learn more when the company is ready.
Do they use adaptive
DSP IP inside their SSDs?
That's a silly question. Of course they do.
In
this quarter - launched today - Skyera's founder Rado Danilak claimed in a
positioning video that his company's vertically integrated technology - which
includes both a new SSD controller and supporting SSD software - can achieve
effectively a 100x gain in endurance using new consumer grade flash.
|
|
| 22 |
Kaminario |
First appearance in the top SSD companies
list.
The list of ingredients on the packet can be a guide to the
product experience - but it's not always as simple as that.
The first
time I read about Kaminario I thought it was
this type of company.
Then I mistakenly thought thought it might be
that type of company.
And it would be an accurate statement now to say that Kaminario are
this type of SSD
company.
I've talked to the company a few times and I'm glad they
remembered that it was me who first suggested they should partner with
Fusion-io.
But
today I think the simplest way to describe what Kaminario is about is to say
that - out of all the rackmount SSD companies - Kaminario is the one which
probably has the best
roadmap symmetry.
(And if I start repeating too many things here which I've already explained
somewhere else - I'm never going to meet my publication deadline. So you'll
just have to click on the link if you want to find out what I mean.)
In
this quarter - Kaminario announced it had secured a $25 million series D round
of funding, bringing its total funding to $65 million. |
|
| 23 |
KingFast |
First appearance in the top SSD companies
list.
In this quarter - some non English language web sites
published reviews of KingFast's (consumer) SSDs. |
|
| 24 |
Micron |
re-entry
In this quarter -
Micron finalized its plans to acquire Elpida
for a
lot of money - which is significant from a memory point of view - but
irrelevant from an SSD viewpoint - because bulk flash memory is just useless
junk unless it's inside an SSD.
What was more significant from an SSD
viewpoint (OK my viewpoint) is that in this quarter I've been hammering on the
walls of Micron's seemingly inpenetrable SSD castle to learn more about the
internal details of their SSDs.
You can read their profile to learn
more. I didn't learn much. But enough.
DSP etc? - Yeah that's a safe
bet. |
|
| 25 |
Toshiba |
Down 5 places since the last quarter.
In
this quarter - Toshiba didn't make any significant SSD announcements.
Toshiba did
complete a transaction with WD related to HDD assets and monopolies
compliance. But that's only HDD
stuff. Part of
storage history
and not very relevant to
SSD's future. |
|
| . |
|
|
 |
| . |
If you've got this far ...
...and
are wondering what to read next...
How about? -
the Top SSD Companies
in Q3 2012 - which was the following quarter.
Or -
the Top SSD Companies
- series overview.
And here are some more suggestions for both very
experienced SSDphiles and newcomers to the market too - in the
SSD articles page.
The
home page of StorageSearch.com
often has content which is closely related to the themes discussed in this
article.
If you're a VP of marketing at an SSD company, or a VC and
would like to know how any SSD company (which is not listed above) may
be doing in search visibility or you want to get my private comments about the
effectiveness of a particular SSD company's marketing, business or technology
plans or have a serious question about anything else which relates to how the
SSD market can get bigger and better faster - just contact me by regular
email using a non
cloaked email and saying what your interest is.
If your question is
interesting enough - or too difficult to find answers for - I'm more likely to
spend time on it.
Finally - if you found this article useful or thought
provoking - please tell other people about it (by whichever method works best
for you).
Thanks. And I hope to see you again soon.
About the publisher -21
years of enterprise guides |
| . |
|
|
 |
