| leading the way to the
new storage frontier||...|
|the top SSD companies
after AFAs -
what's the next box?
can your SSD run backwards?
hold up times in 2.5" military SSDs
where are we
heading with memory intensive systems?
CPUs for use in SSDs in the Post Modernist Era
and risk reward with big memory "flash as RAM"
|rugged SLC like a legendary
blend of Don Quixote and Geronimo has staunchly defied many siren calls to
|Editor:- June 28, 2017 - planar SLC flash may
one day acquire the same romantic mythical status in the SSD market that
tournaments did in the early days of gunpowder. |
Old data warriors
drinking their murky brews on days which blur into each other unpunctuated by
the wake up calls of work related deadlines will say to one another - Dost
thou remember those days when we used to dress up in shiny armor and charge
wildly at each other on horseback while aiming our pointy lances?
had strong muscles and chivalry in those days. Now we just tap an icon to send
a poison cyber capsule or squeeze the virtual trigger for a remote drone
strike... Ah the good old days of SLC (memory you could trust to the care of
childlike simplistic controllers).
For anyone with a grasp of memory
roadmaps and generational progress it has been inevitable for a while that one
day SLC will come to an end - because when the cell sizes become sufficiently
small you need clever controllers to get data integrity. And if you've already
got them for any one purpose then you might as well use them for the next
obvious thing too.
Nevertheless - SLC like a legendary blend of Don
Quixote and Geronimo has staunchly defied many siren calls to surrender and
is still to this day the dominant memory technology in the hotter outposts
of the industrial and military markets.
That's why I draw your
attention to the
by the industrial SSD company Swissbit which
announced that in December 2016, for the first time 1/3 of all the flash
products sold by Swissbit were based on MLC. (Which means there's still time
for your squire to feed dress the horse and polish up the metal jackets.)
Knight of the Seven Kingdoms
- SSD related companies|
|Editor:- June 23, 2017 - I had a quick glance
today at the 2017 Red
Herring North America list to see if I could find companies named there
which intersect with the SSD and memory trends I write about here on the mouse
The strongest match is GridGain
Systems - which I have mentioned
I always set aside time to read their pioneering
in memory computing blogs.
It's not that everything GridGain talks about is new. Many of these
types of applications and problems were being tackled with inhouse developed
code years ago by the early customers of the SSD market pioneers and these
needs fed a lot of startups.
The interesting thing for me about
GridGain - which emerged in 2011 - is they have provided an evangelistic
narrative framework for this aspect of the industry and due to their expertise
they make it sound so easy. GridGain's blogs and customer stories provide a
glimpse of how the whole market will be doing things in the future when
boxes outnumber old style enterprise servers.
TidalScale is a software company
which I keep a regular eye on but haven't written about before.
linkedin ignited interest in TidalScale began in 2016 when Gary Smerdon became
CEO. Gary had previous influential SSD market roles at
software approach to aggregating servers and storage sets a state of what you
can expect to get benchmark model if you're trying to create big memory
from regular good old fashioned hardware. This is without the need to
the servers in hardware using
SCM DIMM wars
and PCIe SSD fabric
For a glimpse at this take a look at their
Performance Gains Without Changing a Line of Code which is one of many
stories related to big memory.
Zadara Storage - which operates
in the storage-as-a-service market - has been keeping my email inbox regularly
updated since 2013 - but Zadara is another company I haven't written about
before because they're not strictly speaking SSD technology originators.
interests are likely much wider ranging than mine so you might be interested in
seeing all the companies
listed in the
|6 years ago
in SSD market
June 2011 -
which soon after became one of the first software companies to break into
the Top SSD Companies
List in Q3 2011
- announced free trials of its software which provided form factor
agnostic tiering, caching and pooling of enterprise flash SSDs. |
many other innovative SSD technology companies which earned fast market
prominence - before and since - it inevitably joined the growing ranks of
acquired SSD companies.
2017 we're now in the
modern memoryfication era when a 4U multipetabyte AFA includes over 250,000
flash memory chips (SSD
news September 2016) so it's not surprising that enterprise users need
software to manage the millions of memory devices which populate a modern
there's a memory shortage?
But assuming you have got those
millions of chips (and remember this is just one cabinet in a datacenter) is it
right that when you have such large semiconductor memory assets that the
pooling model and interfaces still revolve around pretending to be
Or crazier still - SSD arrays pretending to be memory!
why the memory systems market continues to be interesting.
much more you can read about this.
Don't worry about right or wrong
answers. Or not
Whatever you choose you're in the right
place to do your own thinking. We've come a long way since the 1990s and the
world's first mouse-managed
continuously updated resource on mission critical SSDs.
even I have to discard one or two hard learned and cherished SSD ideas which
served me well in previous years but which now have as much relevance as that
rest stop you took when all the roads to where you were heading were closed a
few holidays ago.
Here are some suggested articles.
the SSD heresies
after AFAs -
why DRAM wasn't
good enough to make big memories
software - progress updates
|Editor:- June 7, 2017 - These are some of the
ideas which emerge from a slideshare -
with ReRAM by Crossbar
from a presentation at the recent
- Standard memory busses are too slow to support the computational needs
of new distributed (and always on) AI applications which leverage IoT.
- The only way to improve ultimate "time to get answers"
performance is to integrate storage on the same die as the processor.
- ReRAM can be embedded in SoCs in any CMOS fab to deliver battery
friendly latency under 5nS.
|"The Internet used to
be a one-way flow of information. Systems were only required to enable websites
to provide information feed to the consumer. Today, with more and more
user-generated content and uploads online, storage methods have also become
|Kevin Wang, VP
Sales - Longsys
in a press release -
needs in the IoT and IoV era (June 5, 2017)|
|Toshiba samples QLC says it will sue WDC|
June 28, 2017 - Here are 2 news stories about Toshiba.
yesterday that it is sampling the world's first 3D (64-layer), QLC
(quadruple-level cell) flash memory to
SSD controller makers
The 768Gb chips are believed to be the highest
density nvms currently available from anyone.
Editor's comments:- this
is an extraordinary achievement for the
nvm market. And you can
judge how difficult it has been by comparing the actual timing to the earliest
optimistic market expectations.
In February 2008 - Lane Mason - who
at that time was at Denali
Software and was one of the few people on the planet writing about such
detailed matters - said - in the Denali blog - the industry expected the
transition to 4-bit MLC cells, by 2012. So it has taken 5 years longer. (Lane
is now at Objective
The other Toshiba news story -
Misses its Deadline for a Deal for its Microchip Unit (NY Times) - is
another episode in the much delayed (although not as much as QLC)
of the sale of Toshiba Memory - which many viewers had mistakenly believed
had already publicly aired its season 1 finale. Among other things the NY
Times article says this...
"Toshiba said it was suing Western
Digital for about $1 billion over its attempts to block the sale of the chip
SMART revenue up 38% compared to pre IPO
June 22, 2017 - The first earnings report for SMART Global Holdings
following its recent IPO was
today. Net sales for the quarter were $207 million compared to $149 million in
the same period the year before.
Cactus clarifies role of pSLC etc in embedded SSDs
June 22, 2017 - I had to read a new blog from
twice just to make sure I grasped its meaning properly.
pSLC, MLC and TLC Differences - Does Your Flash Storage SSD Make the Grade?
My did they really say that? moment came when I got
the distinct impression Cactus said something nice about Pseudo SLC.
remember that in the past Cactus has been critical of competitors who offer
pSLC for use in industrial
SSDs - so did I misread something or is it the case that Cactus is now
offering pSLC in its SSDs too?
Yes... to pSLC in Cactus SSDs.
this isn't a reversal of their previous statements.
pSLC is being offered as a memory option for oems who need a high
reliability SSD but don't have the budget for SLC and don't need the full
industrial temperature range either.
Cactus says its "OEM Grade
products are based on pSLC NAND, which is the same MLC NAND as used in
Commercial Grade. The difference is the MLC NAND is set in a mode which only the
top and bottom states are used, thereby cutting the capacity in half but
increase the endurance by 6x the MLC." ...read
Toshiba chooses "paid in Japan" bailout
June 21, 2017 - the long running speculation about
going to buy Toshiba's memory business may soon be over.
its board of directors have resolved to select the consortium of Innovation
Network Corporation of Japan, Bain Capital Private Equity LP, and Development
Bank of Japan a preferred bidder in respect of the sale of Toshiba Memory
Toshiba said its intention is to reach a mutually satisfactory
definitive agreement with the Consortium by the date of its annual ordinary
general meeting of shareholders, scheduled for June 28.
Toshiba's memory business - collected stories
JMR enters the HPC NVMe PCIe SSD market
Editor:- June 16, 2017 - Persistence can be a virtue for
PR folk and not
it's been 15 years since I last wrote about JMR I was still reading
This morning I noticed they're offering NVMe PCIe
SSDs (HHHL and bigger) aimed at the HPC and other high throughput integrator
markets. MSRP of the JMR
SiloStor starts at $795 for 512GB. Capacities available upto 8TB.
the new product matter? Why did I even write this note? It seemed to mark an
important change for me. Click on their profile page link above if you're
curious about my thinking.
BTW if you need any boxes to put JMR's
new SSDs in - they might have some suggestions. They've been in the special
enclosure business since 1982.
Reactive acquires obsolete drive emulation specialist ARRAID
June 16, 2017 - Reactive
the acquisition of ARRAID.
proprietary FPGA-based technology and industrial grade CF cards as the media,
Arraid manufactures new drive replacements for SMD, HISI, HPIB, MAC, OMTI,
Pertec and XMD protocols, and several other legacy drive replacements including
SCSI-1 and SCSI-2 and floppy disk drives. Legacy computer systems simply see
Arraid's drives as if they are the original drives, enabling them to be
field-replaced without the need for any software changes.
HA lessons from BA?
pull any datacenter power cord at
random to attract the attention of a C level BA flight attendant
June 15, 2017 - A systematic failure in fault tolerant architecture and
processes at airline BA
led to hundreds of UK flights being cancelled and delayed over the holiday
weekend at the end of May.
The scale of disruption flashed
headlines in the mainstream tv and news outlets worldwide. But in the days
during and immediately after the story broke it seemed that extracting a
plausible explanation in the public domain was like pulling teeth from the
definitely reputation-damaged and probably litigation-sensitive airline.
was obvious to many experts from the start that failure in fault tolerant
architecture and human error were likely ingredients in the mix. But I thought
I'd wait for a definitive narrative to emerge before placing a note here.
Because it can be useful to learn from the common mode failures of others.
on Bloomberg - Engineer Pulled Wrong Plug provides a good summary
and says (among other things) - "An engineer had disconnected a power
supply at a data center near London's Heathrow airport, causing a surge that
resulted in major damage when it was reconnected."
comments:- to me that's a design architecture fault and shows the failure
to learn any useful lessons from the past 3 decades of enterprise computing
(and in particular the lessons from companies affected by the terrorist
attrocity of 9/11). If BA had a disaster recovery plan it was not fit
backup and disaster
articles from the rotating media era
hard delays from invisibly fixed soft SSD errors can break apps
why you need better storage analytics - says Enmotus
15, 2017 - Using SSDs as its prime example - but with a warning shot towards
the future adoption of NVDIMMs - a new blog -
analytics impact performance and scaling - by Jim O'Reilly - on
the Enmotus blog site -
describes how soft errors can contribute to application failure due to
unexpected sluggish response times even when the data is automatically repaired
by SSD controllers and
when the self-aware status of the SSDs is that they are all working exactly as
That's the needs analysis argument for storage analytics
such as the software from Enmotus
which supports the company's FuzeDrive
Jim says - "Storage analytics gather data on the
fly from a wide list of "virtual sensors" and is able to not only
build a picture of physical storage devices and connections, but also of the
compute instance performances and VLANs in the cluster. This data is
continually crunched looking for aberrant behavior."
Editor's comments:- in my 2012 article -
will SSDs end
bottlenecks? - I said "Bottlenecks in the pure SSD datacenter will be
much more serious than in the HDD world - because responding slowly will be
equivalent to transaction failure."
And in a 2011 article -
the new SSD uncertainty
principle - I shared the (new to me) wisdom collected by long term
reliability studies of enterprise flash done by
STEC - that many
flavors of flash controller management contained within them the seeds of
performance crashes which would only become apparent after years of use as the
data integrity algorithms escalated to progressively more retries and stronger
ECC to deliver reliable data from wearing out (but still usable) flash.
I agree with Jim O'Reilly. You do need more sophisticated datasystems
analytics then whether or not an SSD has failed.
The variable quality
of latency can be a source of
long delays in server DRAM too.
more blogs re 3DX
Editor:- June 6, 2017 - Some
recent blogs about Micron's
- mentions on StorageSearch.com
waiting on 3DXP arrays - by Woody Hutsell suggests
that 3DX is among the aspiring inheritors of the modern
space for latency sensitive customers which historically was once
associated with RAM SSDs.
Woody says - "I think the early usage for 3DXP will flow largely
to server vendors (and their suppliers)."
But he goes on to say
why he thinks that adoption in storage arrays will be slower and at a lower
scale than flash.
despite $27 billion bid - Foxconn considered a longshot
June 5, 2017 - In Apple,
Amazon to Join Foxconn's Toshiba Bid a blog on
EE Times - Jim Handy founder of
Objective Analysis says he
expects the alliances and investing parties may change around before the
final deal is struck.
SSDs - boring right?|
|Editor:- June 13, 2017 - For the past 11 years
one of the safest assumptions you could make about the SSD market was that if
you were looking for excitement and big revenue growth opportunities then the
last place you should be looking was the industrial SSD market.|
an important part of
history is that it started in industrial applications and only became
mainstream and interesting - from a
point of view - when flash SSD designers turned their gaze towards other
directions like the consumer
In fact a good rule of thumb in the exciting days of
disruptive change in the SSD market during 2004 to 2016 was that if you knew the
capabilities of leading edge industrial SSD products in any one year the
picture was probably the same the next year too.
customers - who were concerned about whether they would still be able to get
the same locked BOM
SSD 5 to 10 years after choosing it for their long lived design - the
reassuringly boring predictability and calm, careful approach to new
technology adoption in industrial SSD companies was widely regarded as a virtue
compared to other brasher markets.
But there are many reasons to
believe that the industrial SSD market will soon become a new attention seeking
cauldron of innovation and architecture too.
I describe some of the
indicators which brought me to this surprising conclusion in a new blog -
your grandfather's industrial SSD market - on StorageSearch.com.
|"Order of magnitude
differences between commercial products are rare in computer architecture which
may lead to the TPU becoming an archetype for domain-specific architectures...
Among the success factors of the TPU were the large matrix multiply
(65,536 8 bit systolic MACs) and the substantial software controlled on chip
Performance Analysis of a 92 TOPS Tensor Processing Unit ASIC (pdf) - a
paper by Developers at Google (June 26, 2017)|
- the next box|
| Throughout the
the data storage market we've always expected the capacity of enterprise user
memory systems to be much smaller than the capacity of all the other attached
storage in the same data processing environment. |
new blog on StorageSearch.com
adapted memory systems - asks (among other things) if this will always be
Like many of you - I've been thinking a lot about the
evolution of memory technologies and data architectures in the past year. I
wasn't sure when would be the best time to share my thoughts about this one.
But the timing seems right now. ...read the
there yet? |
| After more than 20 years of writing guides to
the SSD and memory systems market I admit in a new blog on
we there yet? - that when I come to think about it candidly the SSD
industry and my publishing output are both still very much "under
|RAM has changed from being
tied to a physical component to being a virtualized systems software idea - and
the concept of RAM even stretches to a multi-cabinet memory fabric. |
RAM really? - RAM in an SSD context|