| Top
20 SSD OEMs - based on search volume Q1 2011 - ©StorageSearch |
| rank |
company |
editor's comments and recent
milestones |
| 1 |
|
Same as before.
This is
Fusion-io's 9th straight quarter in the #1 slot. Fusion-io's search volume
was 67% higher than the #2 ranked company, and 5x higher than
the #10 ranked company in this list.
Fusion-io has become the
yardstick by which all other enterprise PCIe SSD companies are judged - having
achieved design wins with most of the world's leading server oems and having set
many performance records.
Fusion-io epitomizes what I call the
New Dynasty
architectural trend in the enterprise SSD market. What I mean by that is that
Fusion-io's technology is more suited to new SSD aware server installations and
less suitable as a bolt-on accelerator for legacy setups.
That
leaves a wide door open for other competitors to walk into the legacy side of
the PCIe SSD datacenter - and encouraged by Fusion-io's legitimization of this
interface and form factor more than 30 SSD companies have launched products
in the PCIe SSD market since Fusion-io started to pioneer this segment in
2007.
In
March 2011 - Fusion-io
announced
it has filed a registration statement on Form S-1 with the
SEC for a proposed IPO of shares of its
common stock.
This has focused the attention of many CEOs in the SSD
market to the question of what their own companies might be worth?
If
Fusion-io gets a high valuation then many of its competitors will feel more
comfortable about their own prospects for funding too.
One obvious
question which comes to mind - when you're reading this article is whether
Fusion-io - which is still private company - is worth more than STEC? I guess
we'll know the answer to that soon. |
| 2 |
|
Same as before.
This is STEC's
4th straight quarter at #2. STEC has been in every edition of the top 10 SSD
oems list in the past 4 years. STEC's highest rank was #1 in
2008 Q2 - and
lowest point was #8 in 2009
Q2.
When this series articles started (in 2007) STEC was the
undisputed performance leader in enterprise flash SSDs.
Then things
got complicated as flash SSD architecture broke out of its hard disk form factor
shackles and declared its independence in rackmounts - and was born again in
PCIe cards. Consequently the
fastest SSDs picture
nowadays is blurred - with no single company standing out in the
spectrum of enterprise form factors and interfaces.
That
fragmentation is great if you're the company which "owns" a
particular segment in the minds of customers. It makes your business more
efficient and your messages simpler.
But it's more problematic for
companies like STEC who offer a vast range of products into many diverse
markets which once had few suppliers. In the past STEC could technically sell
SSDs without being good at marketing. But each of those past techie segments
now has 30, 50, or hundreds of competitors. The customers are new too - and
if they see a crowd of products which all look very similar vendors have got to
invest in communications which explain why they think theirs is better.
I've
said many times before on these pages that STEC has particular weaknesses in
the enterprise acceleration market because it does little or no effective
marketing to the end users who actually buy their products and relies instead
on indirect marketing done by its partner oems - who in the main - have lagged
behind the top SSD companies in capturing the market's imagination - and (as we
have been finding out recently in various
Petabyte shipment
announcements) have not been doing as well as they thought in comparative
shipments either.
STEC used to be the company which every other SSD
company wanted to beat - owning (as it did for many years) the
top performance slots
for HDD compatible
server flash SSDs - and a sizable chunk of the
military SSD market
too.
The SSD
market is much bigger now - and the agenda in different market segments is
being set by many different companies - rather than by any single company. If
you try this thought experiment you'll see what I mean. If STEC suddenly
ceased making SSDs - then all its main customers would have little or no
problem adapting their designs to use alternative competing products without
negatively impacting competitive performance. |
| 3 |
|
Same as before.
SandForce has
had strong showings in the top 10 SSD companies lists since the company emerged
from stealth mode in April
2009.
By any measure SandForce today is the best known and most
successful designer of SSD controllers - with over 30 SSD oems using its
technology.
To my way of thinking one of the strengths of the top SSD
oems series is that our search volume methodology was sensitive enough to
predict this success and catapaulted SandForce staright into these listings
in the same quarter as it emerged from stealth mode.
In
February 2011 - SandForce
said it had shipped more than one million of its
SF-1500 and SF-1200 SSD
Processors since they were released into production in 2010.
SandForce
Driven SSD Manufacturers shipped more than 100 Petabytes of NAND flash into
the mainstream computing markets. |
| 4 |
|
Same as before.
Texas Memory
Systems, has been operating continuously in the SSD market longer than any other
company. Despite that the company has often surprised me with its technical and
marketing innovations.
In the rackmount SSD market TMS has established
many performance records with both its RAM SSD and flash SSD product lines -
while its PCIe SLC SSD products - launched in
March 2009 -
remain one of the most popular choices in the market for enterprise users
seeking upgrades for legacy servers.
In January 2011 -
Texas Memory Systems
announced the availability of 8Gbps fibre-channel interfaces for its
RamSan-630 - fast 10TB
3U rackmount SLC SSDs. Each unit can be configured with upto 10 independent
8Gb FC ports for a total data transfer rate of 8 GBytes / sec. Ports can be
mixed - with the previously available (and 25% faster)
InfiniBand. When I
spoke to the company which was traditionally secretive about technical details
I was surprised to learn about some key performance enablers
inside the
RamSan-630. |
| 5 |
|
Same as before.
In the consumer
and SMB market OCZ has been a sales phenomenom. The company has grown its
annual SSD revenue to nearly $200 million within 3 years of entering the
SSD market. Most of that business has been in the highly competitive small form
factor market in which there are now hundreds of alternative suppliers.
One
of OCZ's weaknesses had been that it didn't own any significant SSD IP - which
meant that anything it could do with 3rd party controllers could be quickly and
easily replicated. But in March 2011 -
OCZ fixed that problem by
acquiring SSD controller
company
Indilinx for for
approximately $32 million.
That will enable OCZ to influence future
product features to maximize the fit to user market needs which OCZ has been so
adept at spotting. The newly acquired patent base will also provide horse
trading and licensing revenue opportunities in the long term. |
| 6 |
|
Same as before.
Violin is
another SSD company which shot straight into the top 10 SSD companies
list in the same
quarter as it exited stealth mode in 2007.
Then in
2009 and early
2010
Violin mysteriously seemed to go back into stealth mode. Some speculated that
maybe the product needed refinements - others that the company didn't have the
right business chemistry and resources needed to develop the right customer
base.
In the past year - with new funding Violin is once again
playing the same "very fast" SSD tune to a much more receptive
market audience - now that enterprise users have got more accustomed to SSDs as
being a normal part of everyday business.
In
January 2011 - Violin
Memory
announced
a $35 Million Series B funding round which includes
Toshiba - a strategic
investor since April 2010. |
| 7 |
|
Up 1 place since the last quarter.
RunCore
has more than 50 patents in the field of solid-state storage. The company's SSD
activities started when they developed high speed SSD systems for the military
market in China.
Today RunCore is an international company with a wide
range of SSD modules and cards which span the military, industrial, enterprise
server and prosumer markets. |
| 8 |
|
Down 2 places since the last quarter.
WD's
SiliconDrive SSD family - aimed at high reliability applications in the embedded
systems market - was first launched in 2004.
5 years ago they were the
first company to publicly embark on educating systems designers about the
importance of design differentiators when coping with
sudden SSD
power loss. This is something which more SSD oems have started to talk
about recently as it's a very significant reliability factor.
In March 2011 -
WD announced it
will acquire Hitachi GST
for approximately $4.3 billion. Although the primary motive is
hard drives - the
companies said they would put more resources into SSDs too.
If the
acquisition goes ahead then WD will extend its reach into the enterprise SSD
market initially by means of the
SAS and
FC SSDs which HGST has
recently brought to market.
Reliability has never
been a sexy subject in the storage market. Most users mistakenly believe that
module reliability is similar or good enough from most vendors - and all that's
needed to make it better are some high availability techniques like RAID and
mirroring. But the intrinsic differences between different SSD product families
is huge. And the burden of understanding those differences is just one of the
many facets of SSD
education. |
| 9 |
|
Same as before.
Intel entered
the flash SSD market comparatively late for a semiconductor company - in
2007.
Unfortunately
in the rush to gate crash the
frothing SSD market
bubble, and lacking vital experience Intel has acquired a reputation in
the SSD market for shipping inadequately designed and verified products -
which have required recalls or firmware upgrades.
In the long
term I anticipate Intel will solve its SSD IP problems by acquiring an SSD
company. In the meantime Intel dallies with several
controller companies.
Intel's 510 SSD used a controller from
Marvell and Intel has
invested in Anobit.
In
March 2011 - Intel
published version 1.0 of a new proprietary standard for designers of
PCI SSDs in systems
which use Intel processors - the
NVM Express Optimized PCI
Express SSD Interface.
And
Intel launched a new
2.5" SSD aimed at
legacy notebook
designs which have 3Gbps SATA
ports. The
Intel
SSD 320 (which includes 128 bit encryption) is available with MLC
capacities from 40GB ($89 1k price) to 600GB ($1,069 1k). R/W speeds are
270MB/s and 220MB/s respectively. R/W
IOPS are
39,500 and 23,000. In this new design , Intel has added redundancies that
will help keep user data protected, even in the
event of a
power loss. |
| 10 |
|
Same as before.
Foremay's small
form factor SSDs have appeared from time to time in the
fastest SSDs lists.
Foremay's main markets are in the embedded industrial and military segments.
In
March 2011 - StorageSearch.com
published a new update in the
SSD Bookmarks series
- with
links
suggested by Jack H Winters,
CTO, Foremay.
Foremay announced it is
shipping 32GB PATA
versions of its
OC177
SSD Disk on Chip which measures 22 x 22 x 1.8 mm and has R/W speeds of 70
and 40MB/s respectively. |
| 11 |
|
Up 2 places since the last quarter. This
is Seagate's best ranking in these lists.
Seagate entered the SSD
market later than most other multibillion dollar storage companies in
December 2009.
Seagate has eschewed
acquisition and badge
engineering as
routes into the SSD
market and instead seems to be relying on a combination of in-house
designs mixed with licensing some missing IP.
Seagate has never
appeared in the fastest
SSDs lists.
In March 2011 -
Seagate announced
details of new 2.5"
SAS SSDs - marketed
under its Pulsar brand
- which will ship in the 2nd quarter. Available capacities are 400GB (SLC) and
800GB (MLC). R/W speeds are upto 360MB/s and 300MB/s respectively.
Sustainable random R/W
IOPS are
48K and 22K respectively. |
| 12 |
|
Down 1 place since the last quarter.
Pliant
entered the SSD market in
September 2009
with a family of small form factor enterprise SAS SSDs using its own
controller design. Pliant's route to market is via 3rd party oems who embed its
SSDs into their systems.
Nowadays any competent storage oem can design
its own SAS SSD using a variety of off the shelf
controllers and
interface IP. That puts
pressure on companies like Pliant and STEC to ensure their products maintain a
performance and reliability edge.
As I see it Pliant's current business
model is not sustainable as it has a very narrow channel into the enterprise
SSD market which can easily be choked off by slot substitution. I'm sure that
one escape route out of that trap will be to join the runaway PCIe SSD market.
Another route may be to market integrated storage systems.
(In my 1st
draft of this note I also added at this point - that the company may urgently
seek an acquirer. I toned it down and removed this sentence. But my instinct
was right. In May 2011 - SanDisk
agreed to acquire Pliant.)
In March 2011 - Pliant announced the
retirement of its founding CEO and the appointment of Richard Wilmer
as the new CEO.
The SSD market has changed substantially since
Pliant's original business plans were formed. The company has proved they can
make a fast controller that works. But the SAS market is not the real future
for very high performance SSDs. As I see it the challenge for the new CEO is to
fugure out other attractive ways to leverage Pliant's technology assets and
communicating better with the ultimate customers of their products. |
| 13 |
|
Down 1 place since the last quarter.
SanDisk's best ranking in the top SSD oems list was #1 slot in
Q3 2007.
SanDisk
is one of the largest suppliers of flash storage having shipped 600 million
units in 2010. Most of this is dumb flash storage rather than true SSDs.
SanDisk
sells SSDs both directly to consumers and also via oems (into the mobile
phone, notebook and tv / music player storage markets).
SanDisk is
one of the leading company in advancing the use of
MLC technology in
SSDs, a technology which it inherited from the acquisition of SSD pioneer
M-Systems in 2006.
Despite
occasional talk about "enterprise SSDs" - SanDisk is culturally
rooted in the consumer electronics market. The company has a track record of
preannouncing exciting advanced SSDs or technologies upto more than a year
before they turn into real products (and sometimes much longer than that.)
In
February 2011 - SanDisk
preannounced details of a new
miniature SSD which
will ship in Q3 2011 - the
iNAND
has upto 64GB (x3
MLC) capacity in a 12mm x 16mm x 1mm package. |
| 14 |
|
1st appearance in the top SSD oem
lists.
Kove emerged on the SSD market in
November 2010 -
when it demonstrated a 4U
InfiniBand &
FC compatible
RAM SSD product line
called Xpress Disk (at the
SuperComputing conference) which
can sustain 20GB/s throughput via 6x InfiniBand ports.
At that time
Kove's web site looked like a company still in stealth mode. It's better now
- but still leaves a lot of questions unanswered. |
| 15 |
|
Up 5 places since the last quarter.
LSI entered the SSD
market and sampled its
1st SSD
product in
March 2010.
It was a
PCIe SSD - which is
software compatible with SAS
- an interface which LSI
helped to pioneer.
In 2009 I wrote an
article which
explained why every company in the fast
RAID controller chip
business had to make a long term plan to get into SSDs - because there's no long
term future in pretending to make slow hard drives look fast. The PCIe SSD route
is one of several escape routes for RAID IP companies - and is a natural
direction for LSI.
Building up credibility for a newbie in the SSD
market is tough. LSI has done some of the things in the SSD marketer's ABC
handbook such as sponsoring benchmark reports.
Unfortunately the
early test reports which I saw had been so badly designed that they weren't
useful to any serious performance modeler. This revealed that LSI's marketers
didn't have enough SSD experience to design meaningful benchmarks. That's been
true for many other SSD vendors too.
In
March 2011 -
LSI finally spun off the
Engenio systems
business - selling it for $480 million to
Network Appliance.
LSI had wanted to sell off Engenio since 2004. That should eliminate
conflict of interest issues next time LSI's SSD sales people talk to a
prospective oem customer. And it may put some money in the pot for new business
development activities. |
| 16 |
|
Up 2 places since the last quarter.
Solid
Access Technologies started shipping very fast rackmount RAM SSDs in 2003. The
company has appeared in the
fastest SSDs lists
from time to time.
In recent months I've detected a new note of
confidence among competing RAM
based SSD makers. It's easier to do business - because unlike the old days
- most high end users are now on their 2nd, 3rd or 4th generation SSD
deployments and don't need to be told
what an SSD is.
There are clearer distinctions too - compared to
2007 and
2008 when
flash invaded the datacenter acceleration market - about what the different
roles are for RAM
SSDs and flash SSDs - and why RAM SSDs will always be needed somewhere -
and sometimes is cheaper.
Although Solid Access has been quiet on the
news front for the past several quarters it has benefitted from the rising tide
of interest in this subject. |
| 17 |
|
Up 2 places since the last quarter.
SMART's
SSD heritage goes back more than a decade - and most people familiar with the
company will know of its activities in the embedded industrial and military
markets. Recently the company has been doing more in the enterprise SSD market
too.
In January 2011 SMART published an applications note -
SSD
Power Failure Protection (pdf) which describes the 3 most vulnerable SSD
areas which can get corrupted due to sudden power loss - and describes
typical architectures to prevent it. I only discovered this after I had already
published my own recent
SSD sudden
power loss architecture review article. But this is a very important factor
which differentiates some SSDs from others - even when they use controllers from
the same 3rd party. |
| 18 |
|
Down 4 places since the last quarter.
Samsung stated its aim as long ago as
2005 -
to become the world's largest supplier of
flash SSDs. That
made Samsung the first storage company (of the multi-billion dollar revenue
size) to recognize the strategic importance of SSDs.
Between then
and the 4th quarter of 2010 Samsung's SSD product can best be described as "me-too",
lagging severely behind in the performance dimension, and mostly suited for use
in notebook PCs which the corporation buys you - but which you wouldn't buy for
yourself.
In December 2010 - Samsung started sampling credible 2.5"
SSDs for use in enterprise SATA arrays - and they will probably find homes in
value engineered enterprise storage racks somewhere...
Samsung's real
solution to re-engineering itself as an SSD powerhouse will be to
acquire an SSD
company. It has tried before - in
2008 it tried
to buy SanDisk.
Realistically Samsung needs to buy 5 SSD companies.
- a high IOPS flash SSD specialist for the enterprise acceleration market
- a high reliability medium perfomance flash SSD specialist for the bulk
storage market
- a x3 / x4 MLC specialist for the phone and consumer markets (that was the
thinking behind SanDisk)
- a RAM SSD systems maker
- and an extra one for luck - just in case one of the others goes wrong
In
a couple more years there will be
3x as many SSD
companies in the market as there are in 2011. There will be more choice and
the question of who to buy and why will become clearer. |
| 19 |
|
Down 2 places since the last quarter.
PhotoFast entered the
SSD market in 2008.
The company has appeared in the
top 10 SSD companies
list from time to time.
PhotoFast's products are aimed at the high
performance consumer / prosumer / gaming market.
In February 2011
they launched a 500MB / s 2.5" SATA SSD. |
| 20 |
|
Down 4 places since the last quarter.
EMC
entered the modern SSD market in
January 2008.
No
one who really knew the SSD market was impressed by these products which were
little more than hard disk boxes with a few
STEC SSDs inside.
Compared to native rackmount SSDs at the time they were slow and uninspiring.
Was
that really the best idea that EMC could come up with in response to the gaping
performance gaps in the enterprise market between servers and storage?
EMC
didn't add any worthwhile technology to those early STEC boxes - and by the
time it did start adding tuning
support a few years later - so too were 20 or more other companies.
In
January 2011 - EMC revealed it had shipped 10 petabytes of SSD storage
in 2010.
To put that into context:- it's equivalent to 10%
of the enterprise SSD capacity shipped in the same period by
SandForce
Driven partners and 2/3 of the enterprise SSD capacity shipped by
Fusion-io. Another
difference here is that EMC's SSD sales were to a captive customer market
developed over 2 decades who would probably try any new storage product from EMC
once - whereas the other named companies are still at the early stage of growing
a customer base. |
Just below the surface
The
companies below were near misses to this edition of the top 20 SSD companies
list:- Micron Technology
(21), NextIO (22),
DDRdrive (23),
Memoright (24),
BiTMICRO (25),
Dataram (26) and
Nimbus Data Systems
(27). |
