| leading the way to the
new storage frontier
after AFAs -
what's the next box?
farewell to reassuringly boring industrial SSDs
ideas which became clearer in 2014 /
|Netlist revalidates core
patent related to ULLtraDIMM's core technology|
29, 2014 - Netlist
that the US Patent and Trademark Office has denied petitions requesting
Inter Partes Review (IPR) of Netlist patents asserted against the
petitions were filed by
SanDisk who partnered
with Diablo Technologies
to produce the ULLtraDIMM. Both are co-defendants in Netlist's patent
infringement action pending in the US District Court of the Northern District
of California. Of the 5 petitions filed by SanDisk, 3 were denied in their
entirety and a 4th was denied as to some of the challenged claims.
The PTAB denied institution as to all claims of U.S. Pat. No. 8,516,185 which
relates to the fundamental distributed buffer architecture created by Netlist
and is integral to the architecture of the ULLtraDIMM. Among other functions,
the distributed buffer architecture allows a proprietary DIMM containing various
types of memory to appear to the system as standard DRAM memory.
Netlist's CEO said, "We are very pleased with this outcome and the overall
progress in our multi-year legal proceedings against ULLtraDIMM and SanDisk. The
technologies covered by our patents are critical for the enterprise computing
and storage space. We believe this ruling by the PTAB is a clear validation of
our IP in this area and a testament to the years of seminal development work and
Editor's comments:- One of the things which I had expected to
see in 2014
- but didn't - was the appearance of more vendors competing in the low latency
Although one reason for this no-show could be that
developing any new server based ecosystem requires more software investment
than introducing new products in already established markets - another
(equally important) factor may be that the likely suspects to enter the
future memory channel market are waiting to get clarification from the courts
as to which companies they should be seeking to get patent licenses from (if
Later comments:- yes they were all just waiting. The number of
vendors staking claims to the
SCM SSD DIMM
wars market grew to over 20 companies less than 2 years later.
renewing the spin on MRAM's bright future
December 18, 2014 - Everspin
Technologies said at a recent event they have shipped over 40
million MRAM devices.
one of several interesting observations on the state of the MRAM and RRAM
market contained in a new blog -
Computers that Don't Forget - by Tom Coughlin,
President - Coughlin
Associates - who predicts that the market for MRAM devices may exceed
$2 billion by 2019. ...read
nand flash & other nvm
after they promised to replace flash - it's still here
choice between STTMRAM / MRAM & RRAM will be made by 2017
Western Digital acquires Skyera
15, 2014 -
and HGST today
the acquisition of Skyera.
comments:- This is a momentous acquisition for the enterprise SSD market.
I think the context in which to view this is as the embodiment of a new
wisdom in the industry - that to succeed in the enterprise SSD market today -
and to achieve the ultimate
efficiencies at the
manufacturing level - vendors have to think like systems companies.
some of the biggest systems opportunities for efficient vendors nowadays - in
which efficiencies translate into business opportunities are in hyperscale
systems and internet infrastructure.
I've been discussing these
trends in the past 18 months or so in these articles - which I think are
relevant to today's acquisition announcement.
directions in rackmount SSDs (May 24, 2013) - "One of the most
potentially rewarding market challenges which SSD companies are grappling with
right now is - how to make enterprise solid state storage attractive to users
who aren't worried about their hard drive performance and don't even think they
need SSDs... New SSD thinking inside the box will lead to better enterprise
- meet Ken -
and the enterprise SSD software event horizon (October 8, 2013) - "Why
it's so easy to fall into a trap when answering these questions... How big will
the SSD market will be when SSDs replace hard drives? When will it happen? and
What will be the revenue of the SSD market at that time?"
- Scary Skyera?
(October 22, 2013) - "You may decide that my ratios (of enterprise SSD
capacity needed to replace hard drives) are too timid - I said to Skyera's
founder - if so - scare us!"
- Seagate to
acquire LSI's flash business (May 29, 2014 ) - "...even if Seagate
focused only on the high volume potential of existing cloud infrastructure
customers and big web entities (like Google and Baidu) - who need value based
enterprise SSDs - but who are perfectly capable of designing their own software
and APIs and firmware tweaks - then Seagate could... establish it as one of
(several) leaders in the utility SSD segment of the cloud."
- Skyera's new
skyHawk FS (October 29, 2014) - "Re the mobile data center, new
meanings to unified storage and joining storage gentlemen's clubs..."
Marvell takes a big step into skinny controllers for PCIe SSDs
December 9, 2014 - 3 of the most significant differentiating factors in
flash SSD controller
the past year or so - we've seen a few companies - whose controller products
have long been distinctly at one end of each of the above categories -
introduce new designs which can operate in the other. This is a competitive
response to the realization that any single architecture is better suited for
some applications rather than others and no single design set is best for all
For example - the
SandForce SF-3700 -
launched last year - added big controller features as an option to its legacy
minimalist small architecture configuration mode. And before that - the
intrinsically skinny (cache) SF family added support for optional external DRAM
This week Marvell - whose
controllers have always until now been firmly in the regular RAM flash cache
it now has IP which enables it to play a significant part in skinny RAM flash
cache designs with 2 new controllers:-
- Marvell's 88NV1140 - aimed at
NVMe PCIe Gen3x1 SSDs
- can work with 15/16nm TLC and 3D NAND and needing no external DRAM -
will simplify the design of smaller form factors including the capability of
BGA PCIe SSDs.
- Marvell's 88NV1120 - aimed at the embedded
SATA SSD market which
supports DevSlp - won't by itself create new markets - but will - due to its
small footprint and memory support - lower the cost barriers for greater
adoption of small SATA SSDs in traditional embedded markets.
the Top SSD Companies in Q3 2014
9, 2014 - recently published the 30th quarterly
edition of the
Top SSD Companies -
based on market metrics in
Q3 2014 .
over 5 years - Fusion-io
occupied the top #1 spot in this list as the SSD company which was
most researched by our readers. That was inevitably going to change. And it
did. But not in the way you might have expected.
The new #1 SSD
company is Diablo
Technologies - creator of the Memory Channel Storage platform.
years after the first
year of SSD
revolutions - is the enterprise SSD market about to reignite a new
revolution in server centric SSD architecture?
rise in this list is one indicator that may well be what the market is
thinking about. There are other signs in this market report too.
could say - it's part of a conversation which goes...
What next -
when PCIe SSDs are already everywhere? You need them. But you need more too.
Signs of interesting times ahead in 2015. ...read the article
XES says FPGA based controllers are better for delaying EOL
December 8, 2014 - For SSD specifiers in
- the biggest pressure on BOM stability has always come from obsolete
XES made some interesting
comments about this in a
the SSDs used in the company's rugged flight proven secure SSD based storage
XES says "An important feature of these high-density,
high-performance storage products is that they use an FPGA-based storage
controller. This directly addresses EOL and Obsolescence issues commonly
associated with foreign-designed and manufactured storage controllers, which
often are discontinued before a system can go into production."
DIGITIMES comments on consumer PCIe SSDs
December 3, 2014 - On the theme of
PCIe SSDs everywhere -
a recent speculates that over 40% of new notebooks next year may
have PCIe SSDs inside - compared to under 10% today - due to lower prices (and
standards based controllers).
Kaminario gets another $53 million funding
December 2, 2014 - Kaminario
it has closed an oversubscribed $53 million financing round, bringing total
raised capital to $128 million. Kaminario says it will use the new investment
to accelerate business growth.
Editor's comments:- Kaminario
rose 3 places compared to the previous quarter - in the recently published
Top SSD Companies in
Memblaze uses PMC's NVMe controllers
December 1, 2014 - PMC-Sierra
is utilizing PMC's Flashtec NVMe controllers in its next-generation PBlaze4
Cypress merges with Spansion
Editor:- December 1,
2014 - Cypress
Semiconductor and Spansion today
agreement to merge in an all-stock, tax-free transaction valued at approximately
"This merger represents the combination of 2 smart,
profitable, passionately entrepreneurial companies that are #1 in their
respective memory markets and have successfully diversified into embedded
processing," said ,
Cypress's founding president and CEO. "Our combined company will be a
leading provider of embedded MCUs and specialized memories."
also:- 3 Easy
Ways to Enter the SSD Market
|Megabyte loves reading news|
|In 2014 we witnessed the
birth of a renaissance in SSD inspired enterprise architecture - on a scale of
ambition not seen since the Year of SSD Revolutions in 2007. |
|the Top SSD Companies in
"2015 will be a critical period for NAND Flash manufacturers as they
try to slow capacity expansion."|
Assistant VP -
in his blog (December 4, 2014)|
leverages faster writes within MLC|
|Editor:- December 15, 2014 - One of the core
principles used in
DSP controllers in MLC SSDs is that the less charge you write to a cell -
the less time it takes and the less stress on the flash. From the system
perspective - the result is a faster SSD with lower power consumption and better
Shortening the write pulse means you aren't guaranteed
to easily read back the same data which you wrote. Hence the need for
adaptive DSP based ECC - in which the strength of the ECC used depends on the
health of the memory. And a range of different techniques can be deployed for
different parts of the same memory chip in the SSD.
I discussed the
enterprise market impact of heavyweight adaptive R/W DSP techniques on a company
by company basis in the
2012 Q2 edition of the
Top SSD Companies- and as expected - we saw consequences of having or not
having these technologies in the later competitiveness and acquisition of
several key companies.
Another much simpler variation of these schemes
(from the controller
point of view) - which we've seen used in the industrial SSD market
as - is to redesignate MLC memory blocks as virtual SLC blocks - in order
to get faster speeds and better
without resorting to adaptive DSP.
A new blog from Violin -
Flash - Unpredictable Write Performance - explains the thinking behind an
MLC write scheme which delivers similar speedup results to the industrial
Like them it doesn't require adaptive ECC.
them it conserves the total raw capacity and gets the speedup in a fractional
tier in the memory pool by grouping data into sets according to how much
charge is required to perform subsequent writes.
As the article says
- "It's kind of complicated but potentially we now have the option to
program certain MLC pages using a faster operation, with the trade-off that
other pages will be affected as a result." ...read
Asymmetries in SSD
|Oh flash - how do I love
thee? Let me count the ways|
170 ways for sure. Maybe more
says new SSD market report from Web-Feet Research
|Editor:- December 4, 2014 - In the past 5 years
or so - many of you have heard me talking about "hidden segments in the
enterprise", "defragmentation in the industrial market", "application
speficic SSDs" etc. |
My intention in such articles and in
comments related to SSD news - has been to show you that there are a lot more
uses for specialized flash SSDs then you might think if you started from market
models which segment traditional devices such as hard drives, CPUs and DRAM
and tried to migrate that thinking to SSDs.
That's because SSDs are
not only creating new markets - but - being entirely artificial devices - you
can adapt their characteristics (size, power, performance, reliability, cost) in
any way choose to - if you can find enough customers with needs, budgets and
risk profiles which make doing these design trims worthwhile from a business
point of view.
You might ask - has anybody added these application
I saw a possible answer in an email this morning from Alan Niebel
Alan lists over 170 end-use applications for
flash memory based devices in a new market report -
, CS100FA-2014, (321 pages,
$5.95K) - which also includes many forecasts and projections.
this report - Web-Feet Research provides a density breakout by revenue, units,
and Mbits for SLC NAND, 2-bits per cell NAND, 3-bits/cell and enterprise
2-bits/cell NAND, SPI NAND, Combo (MCP) NAND and 3D NAND; NOR, MLC NOR, serial
NOR, Combo NOR and serial Combo NOR; and Phase Change Memory.
thing you may be sure of is that the 170 figure is just a convenient starting
point to think about the flash market in a way which makes sense in a single
document. But don't be surprised if future editions talk about much bigger
If you think that paying around $5K for a market report sounds
like a lot - just think of how much time it would take you to even make a list
of these different flash products and applications. Sounds cheap when you look
at it that way.
PS - and if you can't find a summary of this latest
report on Alan's web site yet (I couldn't either) - then
also:- market research,
flash memory and nvm