editor - StorageSearch.com
I published the world's first comprehensive history of the SSD market
here (below) on StorageSearch.com which has been used as the primary resource
for many other "so called" SSD history articles in other web sites and
publications - although the attribution to my original article(s) here are
often notable by their absence.
This article lists key technical,
product and market milestones from 1976 upto the
a long read which links to thousands of news stories (mostly written as events
|Coming of Age for SSDs|
decades in "virtual stealth mode", and many false starts and
setbacks, the SSD market has come out as an exciting technology which will
change the way all computers are designed.
Although manufacturers in
the industrial controls market, like Square D and Allen-Bradley were using
rewritable user removable non volatile solid state storage modules as early as
the 1970s, it wasn't till much later that the
evolved into a form which we would recognise today. For most of its early
life, this technology remained an open secret - mainly used in embedded systems
in military applications,
or in high performance
computer research labs.
There were many false starts with
Non Volatile semiconductor
technologies which didn't survive.
In the late 1970s - silicon nitride
EAROMs (electrically alterable ROMs) were marketed by General
Instruments. They had electrically controlled block erase (like later flash
memory). The block erase took 100 milli-seconds using a 42V pulse. Read access
time was 2 microseconds which was only 4x slower than other types of MOS memory
in those days. Unfortunately field use showed that the extrapolated data
life of 10 years wouldn't be achieved in practise. (The erase / write cycle
limit - or endurance as we now call it - was later downgraded by the
manufacturer to 1,000 cycles per cell - for 10 years retention - compared to a
much higher figure in the preliminary datasheets.) As a result many industrial
companies like the company I worked for in 1980 stopped using EAROM and
switched to battery backed CMOS RAM instead.
Dataram sold an SSD
called BULK CORE which attached to minicomputers from
Modular Computer Systems and
emulated hard disks made by DEC and Data General. Each chassis held
8x 256k x 18
RAM modules and had a capacity of 2 megabytes.
- a non volatile RAM card for the Altair (S-100) bus which provided 8KB
storage and cost $650. It used magnetic core store but included all the
interfaces needed to provide compatibility with Intel's 8080 microprocessor.
Strictly speaking (despite having no moving parts) the MM-S100 wasn't an
(definition) - because the memory wasn't implemented by semiconductors.
1978 - StorageTek
the STC 4305 - an SSD aimed at the IBM mainframe
compatible market. The STC 4305 was 7x faster than IBM's 2305 HDD
system (and also about half the price) and StorageTek's SSD - which was
implemented by a "cabinet full" of
coupled devices (pdf) - cost $400,000 for 45MB capacity with throughput
speeds upto 1.5 MB/s.
Moore, who was Manager, WW Systems Engineering for StorageTek at the
time told me (in 2016 for this article) - "The original STK 4305 was built
using CCD. We built less than 25 CCD units before shifting to DRAM. I believe
CCD supplier issues were a key factor for the change. As soon as that change was
made, efforts to add battery backup options gained momentum."
Also in 1978 -
Texas Memory Systems
introduced a 16 kilobyte RAM-based solid state disk system designed to
accelerate field seismic data acquisition for oil companies.
Dataram marketed an
updated version of their
BULK CORE SSD
for use with
DEC PDP-11 and
Data General minis.
In the early 1980s - Intel's 1M bit bubble memory created
a lot excitement as a new non volatile solid state memory technology. Intel
shipped design kits and boards to developers using this technology - which was
positioned as a solid state floppy disk. But it failed to be scalable or cost
effective. Intel spun off the magnetic division in 1987 to
Memtech (who later made
flash SSDs) but bubble memory dropped into oblivion.
In 1982 -
SemiDisk Systems (based in Beaverton Oregon) became the first company to ship
SSD accelerators for the Intel microprocessor based PC market.
first disk emulator cards were S-100 form factor RAM SSDs with 512 kilobytes
capacity ($1,995 price at launch). Internally they used 64Kb DRAM, worked via
a proprietary interface - and were designed to work with
S-100 computers (which had Z80 8
bit CPUs inside with a 64 kilobytes addressible bus). Soon after, SemiDisk
built similar cards for TRS-80 Model II computers, and IBM PC's, and Epson QX-10
computers, and increased the capacity to 2 megabytes when 256Kb DRAM became
Curtis introduced the
ROMDISK SSD for the
introduced SSD storage for the mini-computer market, which was the
hottest part of the server market at that time. EMC's SSDs were 20x
faster than the then available hard disks. But market forces and losses led to
EMC exiting the "memory enhancement" business soon after.
Adtron developed its
first memory card drive for the proprietary memory cards manufactured by Epson
and Mitsubishi. These 1st generation memory card drives found applications with
companies such as GRiD, HP, and Trimble Navigation.
NEC marketed 5.25"
SCSI SSDs using internal battery backed
RAM. (Editor's note - I
benchmarked this SSD on a Sun SPARCengine blade for a real-time embedded defense
app using Oracle. But it wasn't fast enough for what my customer had in mind.)
Equipment Corp marketed the
family of Solid State Disk accelerators for use with its
world's first 2.5" flash SSD to IBM. The 20MB drive oem
price was $1,000
($50K / GB).
Solid Data Systems was
founded. The company soon after patented technology for Direct AddressingTM -
which maximized SSD performance by translating SCSI addresses directly into
DRAM eliminating intermediate delays.
Amperif which had been selling a RAM SSD product called Arctic Fox.
In 1994 - The
SPARC Product Directory
listed 2 SSD products aimed at the Sun server market.
- T8000 - was an 80MB, 10MBps SSD on a single slot
made by Colorado based CERAM. Units in multiple slots could be chained to appear
as a single SSD upto 960M. Performance was 2,000 IOPs.
1995 - EDN magazine called
- "1994's most innovative product for embedded systems."
- SAM-2000 was a rackmount SSD upto 8GB, with 500MBps internal bandwidth-
made by Texas Memory
Systems. The transfer rate through the SBus adapter was 22MBps. Other bus
interfaces included VMEbus
1996 - ATTO
Technology marketed the
II. It was a 5.25" form factor SCSI-3 interface RAM SSD with 64MB
to 1.6GB capacity. Throughput was 80MB/s, and performance was 22,000 IOPS.
1997 - in the SSD market
white paper by Peripheral Concepts listed the main SSD vendors as:-
SEEK Systems, and
Solid Data Systems.
designed a RAM SSD
with hard drive backup. Sales Director - David Trossell told me - "It was
a little ahead of its time and the company dropped it after poor sales."
ComputerSysteme marketed a
of SSD modules which converted flash memory cards into
parallel SCSI flash SSDs.
1998 - STORAGEsearch.com published a daily updated
online directory of solid state
disk vendors - in which Megabyte
was shown chipping away at a rock - which remains the current site metaphor
used for general SSDs.
In 1999 -
BiTMICRO launched an
November 1999 - the number of market active
SSD manufacturers listed on
STORAGEsearch.com had reached 11
2000 - world's first online ads for SSDsIn
- after 8 years featuring editorial about SSDs in our various publications,
Curtis became our first
SSD advertiser. (By
2011 100% of our publishing company's revenue came from SSD ads - as we stopped
accepting ads for any other types of storage products.)
In February 2000 -
unveiled the E-Disk ATE35 - the world's
PATA SSD with 200µS
average access time, 11MB/s burst R/W transfer rates, and 9MB/s sustained
R/W. Capacity options ranged upto 19GB.
- SST (Silicon
Storage Technology) entered the embedded mass data storage market with the
introduction of a flash memory-based ATA-Disk Chip (ADC) product family
packaged in a 32-pin DIP package and are available in a range of capacities upto
64 MByte. SST's ADC used a standard ATA/IDE protocol and could be used as a
replacement for conventional IDE hard disk drives.
- VMIC embedded
Diskonchip SSD into
Linux single board computers.
November 2000 -
Solid Data Systems
published an article - Solid
State File-Caching for Performance and Scalability
- which discussed the declining performance (versus capacity) in new
generations of hard drives - and showed how SSDs could boost the performance of
legacy servers and RAID systems by x4.
Also in November 2000 -
BiTMICRO launched the
world's 1st hot-swappable 3.5" SCSI SSD.
2001 - fast FC SAN SSDs offer 32GB and 100K IOPS in 2UIn January 2001 -
M-Systems sampled the
world's smallest 16MB single-chip
flash disk, the DiskOnChip 2000 TSOP.
In May 2001 -
introduced a product called - FlashSSD - as an option in its OpenRAID
enterprise storage SAN products. This non-volatile solid-state disk was for
the typical 1% to 5% of an application's "hot files" that account for
50% or more of all disk requests. FlashSSD delivered a sustained and
constant 12,000 IOPS and 40MB/s data throughput. The company said - this
data rate is consistent for all types of I/O including all reads and writes, all
sequential and random transactions and for all large and small block transfers.
It could speed up disk based applications by 2x to 5x.
In June 2001 -
Adtron shipped the
world's highest capacity 3.5"
flash SSD. The
S35PC had 14 gigabytes capacity and cost $42,000.
In July 2001 -
Cenatek entered the SSD
market with the launch of its Rocket Drive - a PCI bus RAM SSD which was
designed as a performance accelerator "delivering performance of up to one
million transactions per second." The product's designer Jason Caulkins - went
on later to become the CTO of Dataram's SSD business.
Q1 2001 - SSDs were the 18th most popular subject with our readers.
October 2001 - the number of market active
SSD manufacturers listed on
STORAGEsearch.com had reached 21.
Systems began running ads on StorageSearch.com to promote its
RamSan-210 - which
was a 2U RAM SSD - with 32GB capacity, 4x FC ports, 100,000 IOPS and 20
microseconds access times.
Also in October 2001 -
Adtron - with revenue of
$25 million in 2000 - was named in the Inc 500 list of fastest-growing private
In December 2001 -
Technology announced a channel strategy for its high-performance RAM
SSD accelerator systems to "free applications from the I/O bottlenecks
caused by hard drive-based storage, allowing mission critical files to run from
silicon, rather than from rotating platters."
2002 - 1st NAS flash SSD In
Q1 2002 - SSDs was the 4th most popular subject with our readers.
- M-Systems and
Toshiba announced a
collaboration to market a 16MB version of M-Systems'
DiskOnChip MLC flash SSD
(which later grew to 2GB capacity in 2004.)
|In April 2002 the
banner ad (below) ran here on StorageSearch.com. 2002 was our 3rd year
running SSD ads..|
|In October 2002 -
BiTMICRO set a new
density record with a 77GB dual ported fibre-channel 3.5" flash SSD.|
in October 2002
- M-Systems launched
a patent suit against a company called JMTek LLC - for infringing patents on
USB flash drive technology.
Also in October 2002 -
Michael Hajeck, Dave
Merry and John Conklin
November 2002 -
Gates, talking about Tablet PC's said:- "There are also a lot of
peripherals that need to improve here. ...Eventually even the so-called solid
state disks will come along and not only will we have the mechanical disks going
down to 1.8 inch but some kind of solid state disk in the next three to four
years will be part of different Tablet PCs."
The product shown
on the right - from Imperial
Technology (which is no longer in business) - is an example of a 3.5"
parallel SCSI RAM SSD
featured here on StorageSearch.com in June 2002.
In Q4 2002 -
we ran our first ad for a NAS SSD. It was the
2003 - terabyte SSDs become commercially availableWith the benefit
of hindsight it can now be clearly seen that 2003 marked the beginning of
the Modern Era in the
SSD market (after which date - the only serious technology which could
displace an SSD from its market role was another SSD or SSD related technology).
- Competitors Texas
Memory Systems and Imperial
Technology announced the world's first terabyte class SSD systems.
TMS, provided 2 million IOPS, a 1024 gigabyte capacity, and 128 2-Gbit Fibre
Channel links. It required 2 racks and 5000 watts.
The MegaRam-10000, from Imperial, cost $2 million for a 1TB
subsystem with 48 fibre channel ports.
Q1 2003 - SSDs were 2nd most popular subject with our readers..
launched the WhatsHot SSD analysis tool.
In July 2003
StorageSearch.com announced that "the Solid State Disks (SSD) page was the
number #1 product category out of more than 60 storage focused directories
visited by StorageSearch.com readers in the quarter ending June 30, 2003"
That's why we researched and compiled the world's first annual
State Disks Buyers Guide in July 2003 which collected together in one
convenient document pricing information from across the whole SSD industry.
It covered the range of budgets from under $50 up to $2 million and everything
Also in In July 2003 -
BiTMICRO announced the availability of the E-Disk SafeCapacity software suite
- which identified "hot" files as candidates for migration to the
fast tier of solid
In August 2003 -
Ramtron began sampling
the first FRAM (ferroelectric random access memory) built using 350nm design
rules. The FM25CL04 was a 4-kilobit SPI interface nonvolatile RAM that ran on 3
volts. It could R/W continuously up to 20 MHz with no write delays and was
not restricted by
limits. It offered 10 years data retention and was rated over the industrial
temperature range of 40 to +85 degrees C.
In October 2003 -
Memtech announced that
its Wolverine - a military 5GB 2.5" 9.5mm high PATA SSD designed for use
in submarines, space vehicles and aircraft carriers - was guaranteed to
exceed a minimum of 8 million erase/write cycles.
- SanDisk was
added to the NASDAQ-100 Index
of premier, non-financial, growth companies.
2004 StorageSearch.com conducted the world's 1st survey of SSD Buyer buyer
preferences. We also published the 1st SSD Buyers Guide which included prices,
and the 1st market model estimating the $10 billion / year potential of the SSD
2004 - StorageSearch.com reported that SSDs had become the #1 most
popular topic with our readers in Q1 2004. At that time there was a 2 to 1
difference in capacity between the highest density 3.5" SSD and HDD drives.
- Adtron began shipments
of the industry's first SATA
flash SSDs aimed at the
defense markets. The
A25FB was available with upto 40GB capacity, had 40MB/s sustained R/W speeds,
and included fast secure
erase. Pricing for the 16GB model was $11,200.
In August 2004 -
BiTMICRO launched its
Ace-Disk 2.5" Series PATA SSDs. These rugged
had burst read/write rates of up to 8.3 MB/sec and sustained R/W rates of 2.5
MB/sec and 1.2 MB/sec, respectively. Capacities ranged from 64 MB to 2 GB of
pure solid-state storage.
- BiTMICRO announced
it was developing iSCSI
SSDs. But due to the hyped iSCSI market in 2004 being 10x smaller than
analyst predictions - this product was quietly shelved.
In October 2004 -
Sun Microsystems signed an
agreement to resell rackmount SSD accelerators from
Texas Memory Systems.
Q3 2004 - a solid state disk manufacturer,
Texas Memory Systems,
became the #1 company profile viewed by our readers (out of more than 1,000
storage company profiles in September 2004). We also disclosed that the
Solid state disks directory
(still at #1) got 42% more pageviews than the year ago period.
October 2004 - STORAGEsearch opened the
SSD Survey a 3 month
major market research study to learn more about SSD buyer preferences,
applications and attitudes. Results from the survey were published in articles
in 2005 and detailed findings helped SSD vendors understand the needs of
buyers better, and helped them develop marketing plans which worked around the
prevailing disinhibitors to product take-up and leverage the enablers cited by
buyers in the survey.
Also in October 2004 -
shipped the world's first Ultra320 SCSI flash solid state disk.
- Ramtron announced
that its FRAM had been used in an embedded server card aimed at applications
like gaming machines which required fast and secure data access and where data
integrity is unaffected by power dropouts.
- STORAGEsearch published the 2nd annual
State Disks Buyers Guide. This listed every type of SSD available in the
market by interface type and form factor. It also included a summary of major
developments in the SSD market in the preceding year.|
2004 - It was revealed that Solid State Disks were the Product Category of
the Year 2004 on STORAGEsearch.com based on reader pageviews. The Solid State
Disk page was the #1 category (out of more than 70 vertical storage subjects)
viewed by readers for 44 of the first 50 weeks in 2004. In previous years - the
product category of the year in 2002 and 2003 (2 years running) was SATA. Three
of the world's
growing storage companies in 2004:- (M-Systems, SimpleTech and Texas Memory
Systems) were solid state disks manufacturers.
2005 - Samsung declares SSDs a strategic marketIn January 2005 -
STORAGEsearch disclosed results of the
SSD Survey to
strategic oem customers. The results included buyer preferences for form factor
and interface, budgetary data and factors which would make it easier for SSD
vendors to do more business in future. Selected extracts from the survey results
also appeared in articles and editorial.
In March 2005 -
announced that Bell
Microproducts would distribute its SSD products in North America. This would
greatly simplify the access to this technology for thousands of systems
integrators and oems.
Also in March 2005 -
Curtis (which had
shipped over 15,000 RAM SSDs worldwide) revealed that its Nitro SSDs (3.5" FC RAM SSDs)
had accelerated the network infrastucture in a major phone provider in China.
The customer, the GuangDong Branch of China Mobile was the biggest provincial
branch in the Chinese Telecom industry, with over 30 million users. ...read the case study
In March 2005 - 5 out of the top 10 company
profiles viewed by STORAGEsearch.com readers in March were SSD Makers (out of
more than 1,000 storage company profiles). Site readership grew 6% compared to
the year ago period and pageviews grew by 25%.
Texas Memory Systems
offered the world's first performance related guarantees for SSD products. That
they would outperform any competing storage system, or meet the customer's
agreed application speedup expectation - or the customer would get their money
back. This approach was founded on market research data from
SSD User Survey - which said that users would be more likely to try SSD
systems if vendors offered such guarantees.
Technologies made the first SSD with a
Serial Attached SCSI
published (what turned out to be) a classic white paper -
In May 2005 -
announced it was entering the SSD market with
2.5" drives. This
is the first time in this phase of the SSD market's development that a
multibillion dollar company (Samsung's 2004 revenue was $55.2 billion ) has
entered the market.
Also in May 2005 - this was the first time
that the term "solid state disk" generated enough volume to show up on
the top referring searches to this site.
Also in May 2005 -
volume shipments of the highest capacity
3.5" SSD. The
SCSI - which met MIL-STD-810F
and was rated for industrial
temperature operation - had upto 176GB SLC capacity (in a 1" high
case) and delivered throughput rates upto 320MB/s peak through its
interface. It also supported various
fast purge options.
June 2005 -
availability of the industry's highest capacity 2.5" SATA SSD with 128
gigabytes of storage. SATA had been identified in STORAGEsearch.com's Q404
market research survey as the #1 most popular interface for future applications.
But at this stage in the market's development (Q205) only 10% of SSD vendors
(3) actually offered products with this interface.
2005 - Texas Memory
Systems launched the industry's first SSDs with a 4Gb/s Fibre Channel
interface. The 3U rackmount system offered upto 128-gigabytes capacity and
500,000 random I/Os per second performance.
In August 2005 -
acquisition of one
SSD company by another had (so far) been a rare occurrence but could become
more common in future.
In September 2005 -
SimpleTech launched the
world's first dual interface SSD. At launch time the Zeus Dual Interface SSD,
with both a USB and
offered capacities up to 192GB in a 3.5-inch form factor, and sustained
read/write rates of 60 MBytes per second.
In October 2005 -
Texas Memory Systems
and CCP Games revealed that the world's largest game universe was
accelerated by a 64GB RamSan SSD. A record breaking 17,000+ concurrent
users interacted together within the EVE
Online (science fiction) game environment running on 150 IBM servers. The
SSD resulted in a 40x improvement in performance.
2005 - STORAGEsearch published a new updated market penetration model for
the SSD market called -
Why are Most
Analysts Wrong About Solid State Disks?
Also in November 2005
- Texas Memory Systems
demonstrated the first solid
state disk with a native
at the Supercomputing conference.
Here below is one of the banner ads
we were running in November 2005.
2006 - SSD awareness flares into notebook user marketIn January 2006 -
the first notebook maker to qualify flash SSDs for use in Windows XP, Linux and
Solaris notebooks.The drives used were
in January 2006
announced a new technology called SiSMART built into the company's
entire SiliconDrive product line. By monitoring read/write activity, SiSMART
technology enabled designers of SSD systems to collect raw data from
which to extrapolate flash wear out effects and predict application specific
SSD operating life.
In March 2006 -
started shipping 1.8" 32GB flash SSD drives. Quoting projections from
Samsung said it expected that the SSD market would double to $1.3 billion in
2007 and reach $4.5 billion by 2010.
Also in March 2006 - the
number of market active SSD
manufacturers listed on STORAGEsearch.com
had reached 36.
Technologies became the first SSD manufacturer to display end user pricing
online for the full range of its SSD products. Previously the volatile nature
of memory pricing and fear of price led competition had meant that most
SSD oems declined to publish any pricing data. The SSD pricing exclusion zone
included their own websites, press releases related to product launches, and
even our own SSD
In May 2006 -
Samsung launched the
world's first high volume Windows XP notebook using SSDs.
2006 - SiliconSystems
launched its SiliconDrive Secure family which included the widest range of
security features in a solid state disk.
In July 2006 -
predicted that 50% of mobile computers would use SSDs (instead of
hard disks) by 2013.
in July 2006 - Xiotech
announced support for solid
state disks as accelerators in its Magnitude 3D 3000 virtual storage systems
- making it the first Fibre
channel SAN switch maker to support SSD technology.
August 2006 - the number of market active
SSD manufacturers listed on
STORAGEsearch.com had reached 41.
DV Nation became the
first US reseller to market SSDs online aimed at consumers and SMBs.
- Samsung Electronics
announced first working prototypes of PRAM - Phase-change Random Access
Memory. This is a new non-volatile
RAM technology. Samsung
said PRAM is expected to replace high density NOR
flash within the next
Also in September 2006 - the growth of market interest
in SSDs was revealed by STORAGEsearch.com's
web statistics. Pageviews on our main
SSD page increased 50%
in September compared to the year before period, even though readership had
only grown by 10%. The pageview growth happened despite the fact that the SSD
page had slipped down to #3 (out of hundreds of storage categories.) This
indicates a concentrated shift by readers towards the hottest subjects that
matter most to their future plans. At the same time a greater proportion of the
most popular storage
articles were about SSDs.
Also in September 2006 -
Broadbus was acquired
In October 2006 -
STEC acquired UK
SSD maker Gnutek.
Gnutek's Maracite - a 3.5"
FC flash SSD with R/W IOPS performance of 52K and 18K - provided core
founding IP for what later became STEC's most successful enterprise focused
SSD product line - the ZeusIOPS.
November 2006 - Microsoft
announced business availability of its
Vista operating system - heralded as being the first PC market OS to
include SSD-aware support and native SSD cache management.
November 2006 - SimpleTech
demonstrated the first single chip SSD with
USB or IDE interface. The
chip is available with upto 4GB capacity.
Also in November 2006 -
M-Systems which had
been the fastest
growing storage company in 2004 and a a pioneer in the use of MLC in SSDs.
In December 2006 -
PC Accelerators - which described in detail how the recently launched
Windows Vista OS supports solid state disks.
Also in December
2006 - Advanced
Media entered the SSD market taking the total number of SSD manufacturers
listed on STORAGEsearch.com to 44 - which is 4 times as many as in 1999.
I called 2007 - the "Year of SSD Revolutions".
This was the year in which
2.5" and 3.5"
flash SSDs from Mtron
and Memoright broke
away from the me-too performance pack - and showed that single flash SSD
drives in traditional HDD form factors could economically challenge the R/W
throughput and random IOPs of the fastest enterprise
SSDs from EasyCo (array
of COTS SSDs) and Texas
Memory Systems (proprietary flash array) showed that flash SSDs could
replace some market niches previously held by
RAM SSDs - at much
lower cost and without worrying about wear-out.
For all SSD
milestones in 2007 - click here. Below is just a partial list.
February 2007 - amid competing claims from various other oems
Mtron launched the
fastest 2.5" PATA SSD - with 80M bytes / sec sustained write time.
March 2007 - SanDisk
joined the overheating market for 2.5" SATA SSDs.
2007 - Fujitsu
announced it had terminated plans to manufacture 1.8"
hard drives for
portable products - because in this form factor it said
SSDs can offer better
speed, lower power, lower weight and lower cost.
In May 2007
announced its new flash chip technology could deliver 800M bytes / second
sustained throughput for flash SSDs using today's technology.
2007 - SiliconSystems
said that it had received an additional patent for its PowerArmor voltage
detection and regulation technology. PowerArmor, used in the company's
SiliconDrives protects critical operating system files and application data
from corruption due to
In August 2007 -
Violin Memory launched
world's fastest 2U SSD.
In September 2007 -
Texas Memory Systems
launched the RamSan-500
- which delivers 2 terabytes of high speed flash SSD in a 4U rackmount package.
Performance is 100,000 IOPS sustained random read, 10,000 IOPS sustained random
write. Throughput performance from fibre-channel hosts to internal flash storage
is 2G bytes / sec sustainable (3G bytes / sec peak).
In October 2007 -
itself on the international SSD market when they demonstrated a range of 2.5"
64GB PATA and SATA flash SSDs which were the
fastest in the
market at that time - with 100MBytes/sec sustained read and write transfer
In November 2007 -
Micron Technology said
it would launch a family of SATA 1.8" and 2.5" flash SSDs in Q1 2008
bringing the total number of market active
SSD oems to 60.
December 2007 -
SSD Alliance is
founded to develop compatibility standards for flash SSDs.
Also in December 2007 -
Toshiba said it will enter
the SSD market with 1.8" and 2.5" SATA models which will be sampled in
Year of the SSD Centurians. This is the year which the
number of SSD oems passed
100 companies, and in the server market fast flash SSDs broke the
asymmetric R/W IOPS
all SSD milestones in 2008 - click here. Below is just a partial list.
- After a 20 year gap
EMC re-entered the SSD
market with the launch of its Symmetrix DMX-4 networked storage systems
populated with SSDs from
In February 2008
- SMART acquired
Adtron for a sum in the
region of $20 to $35 million.
In March 2008 -
OCZ entered the SSD
market with a 2.5"
flash SSD - taking the number of
SSD oems listed on
STORAGEsearch.com to 70.
In April 2008 -
Seagate filed suit
STEC alleging patent
infringements related to hard disk interfaces. The case was seen by many SSD
proponents as a potentially deadly but seriously misguided missile launched
at the entire SSD market. It was later dismissed without merit.
May 2008 - California based
opened its first office in the People's Republic of China.
2008 - Fusion-io
said it was adapting its flash SSDs (click to see pdf)
to provide acceleration in HP's BladeSystem servers.
In July 2008 -
STORAGEsearch.com published an
Can you trust flash SSD
specs & benchmarks? as a warning to readers about the unreliability of
many SSD test results which were being published on the web which we knew
weren't set up properly.
In August 2008 -
Violin Memory said it
had delivered 1 million IOPS on a single interface port (a world record)
using the latest version of its Violin 1010 memory appliance. Violin also said
that its new technology would deliver 100K write IOPS on a future flash SSD
version of their product.
In September 2008 -
Samsung published an
letter aimed at shareholders offering to buy
SanDisk in an effort to
subvert SanDisk's rejections of its earlier attempts to acquire their company
and flash IP.
In October 2008 -
Intel started shipping
the X-25E - a
flash SSD. Read
latency is 75 microseconds and a 10 parallel channel architecture enables it to
sustain R/W throughputs of 250 / 170 MB/s. Random IOPS performance is
impressive with a 10 to 1 R/W ratio which is inline with the best
designed enterprise flash SSDs. Using 4kB blocks - random R/W IOPS are 35,000
and 3,300 respectively.
Also in October 2008
IMEC said it had
new research activities on resistive RAM (RRAM) cells. These research
activities were focused on non-volatile memory (NVM) technology, concepts and
solutions for the 32nm generation and below. At the time RRAM was described by
IMEC as a "potential candidate to replace conventional
flash memory below the
22nm technology node".
In November 2008 -
Violin Memory announced
availability of a new 1010 Memory Appliance - a fast 4TB SLC flash SSD in a
2U rackmount. Its patent pending non blocking architecture delivers the best
ratio of flash R/W IOPS in the industry - over 200K random Read IOPS and 100K
random Write IOPS (4K block). Interface options include:-
Fibre Channel and
Also in November 2008 - the world's first online ads for PCIe
SSDs started running here on the mouse site. Below you can see the 1st draft "temporary"
banner ad which was seen by hundreds of thousands of readers. Fusion continued
advertising their PCIe SSDs on StorageSearch.com right up to the time they
were acquired 6 years later.
published a significant whitepaper -
Evolution and its Effects on SSD Useable Life (pdf). Starting with a tour
of the state of the art in the flash SSD market the paper introduces several
new concepts (including write amplification and wear leveling efficiency) to
help systems designers understand why current wear usage models don't give a
I explained why I thought 2009 would go down in history as
the Year of SSD
Market Confusion. This is the year in which search volume for
surpassed that for any other SSD form factor - knocking
2.5" SSDs off the
It was also the year that flash SSDs reached the same storage
density as hard drives in the same form factor.
For all SSD
milestones in 2009 - click here. Below is just a partial list.
- pureSilicon said it is
sampling the highest density
2.5" SSD - with
1TB capacity in a 9.5mm high form factor. Sustained read / write performance is
240MB/s and 215MB/s respectively. The
SATA SSD has latency
under 100 µsec and is rated at 50,000 read IOPS, and 10,000 write IOPS.
I asked if compression was involved in achieving the capacity - but was told -
no. Internally it's got 128 pieces of 64Gb MLC NAND.
In February 2009 -
became Chief Scientist at
March 2009 - SiliconSystems
announced that it has shipped over 4 million SiliconDrives integrated
with the company's
In April 2009 -
SF-1000 family of SSD
Processors - aimed at oems building SATA flash SSDs. Its 2.5" SSD reference
design kit is the fastest 2.5" SATA flash SSD on the market - with 250MB/s
symmetric R/W throughput and 30,000 R/W IOPS.
Fusion-io was named the
#1 company in StorageSearch.com's
list of the the Top 10
SSD OEMs based on search volume in Q1 2009. This was the 1st time that
the #1 slot has been held by a company which does not make traditional
In May 2009 -
DDRdrive emerged from
stealth mode and launched the
DDRdrive X1 - a
RAM SSD with onboard
flash backup. Load / restore time is 60S. Performance is over 200K IOPS (512B).
R/W throughput is 215MB/s and 155MB/s respectively. Capacity is 4GB. OS
compatibility:- Microsoft Windows (various). Price is $1,495.
2009 - DTS won a
of show award at Interop
Tokyo 2009 for its Platinum SSD. The company says it will ship a 2.5"
version of this product - which
delivers about 40,000 IOPS and 250MB/s R/W - later this month.
July 2009 - STEC
announced it had received
million order for its
ZeusIOPS SSDs from
a single enterprise storage customer (later confirmed to be
EMC) for delivery in the
2nd 1/2 of 2009.
In September 2009 -
Intel said it will
deploy up to 10,000 SSD notebooks this year to its own employees following an
review of the benefits.
In October 2009 -
it has started sampling its 1st SSD product to major oems.
2009 - Fusion-io
details of a very fast PCIe form factor,
compatible, flash SSD designed for 2 undisclosed government customers. Each
ioDrive Octal card, occupies 2 slots and delivers 800,000 IOPS (4k packet
size), 6GB/s bandwidth and has upto 5TB maximum capacity (implemented by 8x
In December 2009 -
Micron announced it is
SATA MLC SSDs in 1.8"
and 2.5" form
factors. Micron's C300 SSD can achieve a read throughput speed of up to 355MB/s
and a write throughput up to 215MB/s.
As 2010 was about to dawn I explained why I thought this would be seen as
the start of the SSD
market bubble. This was the first year that SSD market revenue reached
billions of dollars.
quarter - among other things...
- Fusion-io (the
#1 company in the quarterly
top 10 SSD companies
list) announced 300% annual revenue growth - thereby confirming
that its New Dynasty
approach to the market was getting a positive reaction to those with budgets and
the company was not simply a favorite with editors and
quarter - among other things...
- ioSafe launched the
ioSafe Solo SSD - an ultra rugged USB
/ eSATA external flash SSD with upto 256GB capacity ($1,250) designed to
provide data protection against disasters such as fire, flood, and building
collapse. At this stage of the SSD market development less than 3 companies had
talked seriously about the subject of
using SSDs as backup. It
would be several years later before that market emerged with a strong identity
and became a billion dollar market in its own right.
2010 3rd quarter
- among other things...
announced it is sampling
SSDs based on its patented Memory Signal Processing technology which provide
20x improvement in operational life for MLC SSDs in high IOPS server
environments. This guarantees drive
of 10 full disk writes per day, for 5 years.
- Foremay announced
it is shipping 2TB 3.5"
and 1TB 2.5"
SATA flash SSDs in its EC188 M-series model V product range. R/W speeds are up
to 200MB/s. ECC is 24-bit. The SSDs are bootable and support all major
- NVELO launched
Dataplex - a software product
aimed at PC oems - which provides
functionality inside a
Dataplex said it will begin shipping from select Tier 1 PC OEMs in 2011.
2010 4th quarter
- among other things...
that NDS (a tv set top box designer with
with over 30 million DVR units deployed) has successfully has designed SanDisk
SSDs into a new range of lower cost set-top DVRs. SanDisk asserted that SSDs
are cheaper than HDDs in entry level DVRs
- Samsung said it
is shipping 200GB 3.5"
SATA SLC SSDs to EMC.
Sequential R/W speeds are 260MB/s and 245MB/s respectively. R/W
47,000 and 29,000. The new Samsung SSDs have an 'end-to-end
function and encryption.
Business activity in the SSD
market was energized by the realization that SSD companies were worth a lot of
money. Initially indicated by the valuation of
Fusion-io's IPO in
the 1st half of the year - a spate of acquisitions of SSD controller companies
later in the year revealed that the storage industry had great expectations
for the future size of the SSD market.
among other things... here were
the main highlights
- March 2011
Indilinx for for
approximately $32 million.
WD announced it
will acquire Hitachi GST
for approximately $4.3 billion.
it had filed with the SEC for a proposed IPO. When this went ahead - in the
following quarter - the company's market cap was nearly $2.5 billion. That
created a lot of excitement and confidence in other enterprise SSD companies
who were thinking about what their own private companies might be worth.
- June 2011
- FlashSoft launched
its first product - software which enables enterprise flash to be used as a
cost-effective, server-tier computing resource (ASAP functionality in
software) which is available for free evaluation through a 30-day "Try
Before You Buy" program.
- July 2011
- OCZ is sampling a new
dual core ARM based
SSD controller for
6Gbps SATA SSDs which
can deliver upto 500MB/s sequential throughput and 200 mega transfers per
platform supports up to 1x nm NAND Flash with 1, 2, or 3 bits per cell,
has 70 bits of BCH ECC
per sector, end to end data protection, fast boot options (50% faster than
competing SSDs) and enhanced
protection. The new platform - supports 1TB flash capacity and has a
400MHz DDR3 DRAM
cache interface with support for up to 512MB.
announced it is
sampling the fastest (yet) 2.5"
SATA SSDs based on its
own controller design. The new
SSDs (upto 400GB) delivers up to 70,000/40,000
block size) and 510 MB/s sequential read/write with non-compressible data
using 2xnm MLC NAND. Anobit says its patented Memory Signal Processing
from 3,000 write cycles to over 50,000.
I anticipated 2012 would be
the year of the enterprise
SSD market goldrush but 2012 turned out to be more complicated than
that - being at the pivotal point in several long term multi-year
Transitions in SSD.
- October 2011
a definitive agreement to acquire
approximately $370 million.
Viking shipped the
industry's 1st flash backed hybrid DRAM DIMMs. The DDR3 ArxCis-NV plugged
into standard RAM sockets and provided 8GB RAM which was backed up to SLC flash
in the event of a power failure - while the memory power was held up by an
external 25F supercap pack. Viking said these new memory modules eliminated the
need for battery backup units in servers and the maintenance logistics
associated with maintaining them.
- April 2012 - DensBits
sampled a new SSD controller - the (adaptive R/W DSP based) DB3610 -
which supported the latest 2Xnm and 1Xnm TLC (3 bits/cell ) MLC flash with
an extreme endurance
figure of more than 10K P/E cycles and R/W performance of up to 95MB/s / 65MB/s
and 4,000 / 1,100 R/W
for sequential and random operations, respectively.
SMART announced that
it was able to get 5x more endurance using unmodified industry
standard controllers from LSI/SandForce - by
preconditioning the flash memory based on better write pulse parameters learned
from its own Optimus DSP based controller.
- August 2012 - IBM
launched its first product - a 1U half-depth 10GbE
SSD rack with 44TB usable capacity costing $131,000 approx - which used
adaptive R/W flash IP.
- September 2012 - 3 of the top 10 SSD companies (Virident,
STEC) announced new
CEOs for a variety of reasons.
- October 2012 - Virtium
outlined a new 4 part categorization scheme for matching embedded
temperature rated SSDs to user cases and needs.
OCZ slimmed down its catalog
and headcount - by 150 products and about 30% respectively - to re-engineer
Proton Digital Systems
emerged a bit more from stealth mode.
2013 - a
new threat and new roles for PCIe SSDsIn the 6 years leading up to 2013 -
the embodiment of what was the fastest type of flash SSD which could cozy upto
apps servers plugged into enterprise CPU motherboards had been
PCIe SSDs. There had
been indications in 2012 that PCIe SSDs could reach out and do more within the
enterprise - and in 2013 we witnessed some of those anticipated changes:- such
as the wider software
support for "beyond the same box" clustering and fabrics for PCIe
SSDs - which simplified
and began to encroach on SAN-like
functionality - albeit at a much reduced - cabinet-like physical distance. The
market for 2.5"
PCIe compatible SSDs - also became served by more vendors.
role for "slower" PCIe SSDs also began - in the shape of products
designed for consumer PCs and ultrabooks. These offered latency and throughput
which was vastly superior to anything which had previously been available in
mainstream notebooks - which had been hampered by an over reliance on
fossilized hard drive interfaces such as
competitive pressure in the fast PCIe SSD market space also emerged this year -
with the debut of a new class of motherboard SSDs designed to provide flash
style capacity but designed to operate transparently within standard DRAM
sockets - memory
For the SSD market generally one of the
changes in 2013 was that - unlike before - new companies entering the
market could rely on a sophisticated SSD ecosystem - in which key elements of
their solutions were already being supplied by other companies.
are some of the highlights and milestones for 2013 extracted from the archived
SSD news for each month.
In January 2013 -
BiTMICRO launched a
marketing program to license its Talino SSD controller.
Foremay shipped 2TB
SATA SSDs with standard 9.5mm thickness.
In February 2013 -
announced beta availability of a new software suite - called FlashMAX Connect
- which enables a single PCIe flash card - made by the company - to service
$51 million in financing
Dell Ventures and including an
investment by WD.
- EMC said it was sampling
flash arrays which are designed and managed using the
controller architecture based on leveraging IP from its
- Diablo Technologies
SMART Storage as
its exclusive flash partner to pioneer a new type of (faster than
NexGen Storage (an
hybrid array IP
company) for $119 million.
In May 2013 -
that its co-founders - David Flynn
(who had been CEO and President) and Rick White (who had
been CMO) have resigned and will pursue future entrepreneurial investing
it's sampling a new model in the
hot swappable 2.5"
PCIe SSDs market - the
has upto 1.4TB MLC capacity and can deliver 750K R IOPS. Micron specifies
endurance as "50PB of drive life".
In June 2013 -
WD announced that it
had agreed to acquire
Stec for approximately
$340 million. Stec will be absorbed into
Samsung entered the
PCIe SSD market - with
models aimed at notebooks.
In July 2013 -
a definitive agreement to acquire SMART Storage Systems
for approximately $307 million.
announced its entry into the
2.5" PCIe SSD
market. Its NVMe SSD had upto 1.6TB capacity, read throughput upto 3GB/s,
and up to 740K
SMART Storage Systems
it had begun sampling the first
SSDs compatible with the interface and reference architecture created
Technologies. SMART's first generation enterprise
(ULL = ultra-low latency) can be deployed via any existing DIMM slot and
provides 200GB or 400GB of enterprise class flash SSD memory with upto 1GB/s and
760MB/s of sustained read/write performance, with 5 microseconds write latency.
it had closed an oversubscribed $150 million Series E funding round with
institutional investors which brought the company's total capital raised to
In September 2013 -
sampling the first implementation of the
Hybrid Memory Cube. Micron's
(short reach) HMC provided 2GB DRAM in a BGA - with upto 160GB/s bandwidth.
began trading on the NYSE Friday as "VMEM". The company got the
money it wanted - $162 million - but those who bought at the original price
didn't consider themselves quite so lucky.
Insider commented on Violin's Awful IPO
In October 2013 -
it has licensed its advanced Memory Modem technology (a variety of
and DSP flash controller IP) to
Toshiba for use in
new designs of SSDs.
In November 2013 -
3rd generation SandForce SSD
controller family - the SF3700 - which offered native SATA or gen 2
PCIe interfaces - and incorporated adaptive R/W DSP ECC management.
StorageSearch.com described the
underlying business concept as "SSD market on a chip" - and
described the design as the most ambitious design of a single chip
SSD controller in
Primary Data -
which had been founded less than 6 month earlier by the departure of the
two cofounders of Fusion-io
- announced it had secured a $50 million funding round.
Maxta launched its first
product - the Maxta Storage Platform
- a hypervisor-agnostic software platform for repurposing arrays of standard
servers (populated with cheap standard
SATA SSDs and
hard drives) into
scalable enterprise class apps servers in which the global CPU and storage
assets become available as an easily managed meta resource with optimized
performance, cost and resilience.
In December 2013 -
that it had agreed to be
acquired by Avago Technologies Limited in an all-cash
transaction valued at $6.6 billion.
2014 - In
we witnessed the birth of a renaissance in SSD inspired enterprise
architecture - on a scale of ambition we haven't seen since the
Year of SSD
Revolutions in 2007see
SSD ideas in 2014
IBM disclosed it had
shipped over 1,500 fast
rackmount SSDs in
its FlashSystem 800 range - based on the RamSan product line acquired from
Texas Memory Systems.
IBM also launched a new range of servers which were the first in the market to
include - as standard -
SSDs (low latency DDR3 DIMM flash SSDs) supplied by
SanDisk and based on
the design created by SMART
- A3CUBE unveiled a
new PCIe memory fabric for 10,000 node-class architectures. The
(pdf) remote shared broadcast memory network provided 700nS raw latency
(4 byte message) and promised message throughput - via standard PCIe -
which would be 8x better than
In March 2014
it was sampling a new
5.0 controller - the 88NV1088 - thereby enabling "SSD class"
performance (280MB/s read speed and 5K random IOPS) in a smartphone compatible
- VMware enterd the
SSD software market with the launch of VSAN version 1.0 (an SSD ASAP / hybrid
virtualizing appliance) - which supports 3-8 server nodes.
In April 2014
- Coho Data
announced general availability of its first product - a 2U SSD ASAP called the
DataStream - which integrated
hard drives and a
server into a web scale expandable unit (using an internal 52 port 10GbE fabric
switch) to implement what the company refers to as a "MicroArray"
with 180K IOPS performance (4KB).
In May 2014 - Seagate agreed to acquire
business for $450 million.
- SanDisk began
sampling 4TB SAS SSDs
in a 2.5" form factor. The Optimus MAX was rated at 1-3
In June 2014 - SanDisk
launched a new
enterprise software product -
ZetaScale - designed
to support large inmemory intensice applications. ZetaScale's technical roots
came from the earlier acquisition of Schooner. At the time of the launch -
StorageSearch.com said that ZetaScale was likely to be one of the most
significant SSD software products launched in 2014.
In July 2014 -
NxGn Data exited stealth
mode - promising it would sample M.2 form factor SSDs for the enterprise
market with in-situ SSD processing APIs in 2015.
In August 2014
announced details of a new 2nd generation
SSD - low latency flash SSD accelerators in DDR-4 sockets - which will
sample in the first half of 2015.
- Tanisys Technology
entered the SSD testing market
- with a product line of burn-in chambers and performance analyzers which
included adapters and fixtures for many of the popular SSD interfaces in the
In September 2014
- Shannon Systems
made its first appearance in the US market at the
Flash Memory Summit. The
company described itself as - "the leading provider of PCIe flash and
associated flash systems in the China market."
- A3CUBE said that
the first US customer shipments would occur soon of its PCIe connected
shared reflective memory fabric - RONNIEE Express
In October 2014
reported that industry adoption of its
(flash+HDD) for the consumer PC market in the past 2 years had been
significantly lower than expected.
- Netlist asked a US
court to shut down production of SanDisk's ULLtraDIMM
(memory channel SSDs) and recall units sold - as a step in its long running
patent dispute with Diablo.
In November 2014
said size matters in markets such as the mobile data center - at the
launch of its 136TB (raw) 1U rackmount SSD - the skyHawk FS.
- Steve Wozniak
joined Primary Data
as Chief Scientist - heralding the company's exit from stealth mode.
Foremay said it was
accepting orders for a high capacity - 8TB - 2.5"
military SATA SSD.
Mass production was planned for Q1 2015.
- Western Digital
(which at that time was offering the highest capacity SSD racks based on its own
controller architecture and software).
- Marvell sampled a
new range of SSD controllers which would support skinny (RAMless) NVMe PCIe
Gen3x1 PCIe SSDs working with with 15/16nm TLC and 3D NAND in BGA PCIe SSDs.
In January 2015
- Tezzaron said it
would use Rambus's ReRAM
technology in production SSDs in 2016.
- SanDisk spun off
NexGen Storage as a
separate company. This was - in effect - undoing the last acqusition made by
Fusion-io in the year before its acquisition by SanDisk. SanDisk cited as a key
reason being it didn't want to remain involved in an enterprise storage
business which included hard
In February 2015
acquired the HLNAND related technologies, assets and patents of
known as MOSAID.
announced a new SSDcentric storage pool redeployment and management platform
called FreeStor - which enables users to support many different flash array
brands using a single set of abstraction tools.
- OCZ announced
a collaboration with Levyx to develop and
validate a flash as DRAM solution as a competitive alternative to
DRAM rich server arrays
for big-data real-time analytics.
- Waitan launched a
2.5" SATA SSD with 4TB capacity with special security options to protect
purge data if the SSD
gets into the wrong hands. The StellaHunter - aimed at UAV, drone and similar
military applications - can use 2 or more predetermined zoning conditions to
initiate self destruct - unlike previous autonomous self destruct SSDs which
acted on a single trigger which could be prone to false positives.
- Toshiba began
sampling 48-layer 3D stacked nand chips with 16GB capacity.
In April 2015
- Diablo won a "decisive
victory" in its critical patent based court battles with
Netlist related to
SSDs (ultrafast flash DIMMs). This meant that production - which had been
suspended due to an injunction - was able to resume.
In May 2015
- WD disclosed that
its enterprise SSD revenue had grown 67% yoy to $224 million / qtr.
This is a revenue run rate of $1B / year in enterprise flash.
- Nimble Storage
that revenue in its recent quarter grew 53% yoy to $71 million.
- A notable feature reported in
OCZ's dual port 2.5"
NVMe SSDs was user-selectable power envelopes, in 15W, 20W and 25W settings.
- Altera introduced
adaptive DSP flash controller IP into its FPGA product line.Nantero received $31
million funding for its high temperature 300º C retention nvram.
- Savage IO
launched a 4U server storage box with "more lanes of SAS than anyone else"
- Intel and
Micron began the
(technical details devoid) unveiling and pre-announcements of a new jointly
developed non flash NVDIMM product family called at various times in the
following weeks - "3D XPoint" and "Optane".
- Aupera launched the
world's first M.2 SSD which
used MRAM (instead of flash) as the
In September 2015
- Diablo launched a
new DIMM compatible platform called
designed to replace DRAM in big memory servers. At the time of the launch
StorageSearch.com said - "Memory1 is the most significant development in
the enterprise flash market in the past 3 years."
got $10 million series B funding for fastest NVMe flash.
got $35 million series D funding to make RRAM SSDs a reality in 2016.
- Dell agreed to
acquire EMC for $67
In December 2015
announced it was working with Samsung to develop
The phoney war in DIMM wars ended with real products
shipping. Big memory ideas about tiering and SSD management migrated from
webscale all the way down to embedded low power products. It still wasn't clear
which memory technologies would win the SCM DIMM wars but with a growing
software ecosystem the risk of choosing the wrong hardware product didn't matter
announced the beta release of Software Defined Memory.
- Quarch Technology
said that many enterprise SSDs still failed stone dead after hot swap tests.
- A study in the field reliability of PCIe SSDs by Google concluded that
it wasn't worth paying more for SLC than MLC.
In March 2016
- The industrial programmable controller maker Schneider said it had adopted
Everspin's MRAM for
use in future designs.
In April 2016
- NVMdurance got
a new round of funding for its endurance stretching technology.
- Samsung began
mass producing the industry's first 10nm class 8Gb DDR4
In June 2016
- Symbolic IO unveiled a new system
level, memory intensive enterprise architecture which used efficient data
packeting to reduce power and accelerated performance.
- Pure Storage
said its AFA revenue in Q1 2016 was more than the leading HDD array brand.
- Diablo announced
volume availability of its Memory1 128GB DDR4 system memory module.
- Seagate previewed
a 60TB 3.5" SAS SSD.
In October 2016
- Everspin filed
its IPO to get funding to expand the use of MRAM
In November 2016
- Rambus announced
it was exploring the use of Xilinx FPGAs
in its Smart Data Acceleration research program.
- Silicon Motion
announced the "world's first merchant SD 5.1 controller solution"
ained at the Android phone market.
- Violin said it was
seeking bankruptcy protection.
In the first half of 2017 there
were 2 background factors which surfaced in many stories related to the SSD
- The memory supply shortage (due partly to yield problems with new
generations of 3D nand) affected the whole market.
For example Seagate
said it had lost $50 million of SSD business in a single quarter due to
unavailability of raw memory.
Looking back on 2017
as a whole it was the best of times for some and the worst of times for
- The risk of bankruptcy for Toshiba's parent company due to losses in a
This gave rise to much
about who would be the successful bidder for the company's memory and SSD
business - which Toshiba said it would sell to fix its balance sheet.
In January 2017
announced it was sampling 8Mb ReRAM based on 40nm CMOS friendly technology.
In March 2017
- Pure Storage
said the "new stack" is becoming the standard thing.
began sampling an NVMe PCIe SSD based on its ST-MRAM.
In April 2017
- Intel began
sampling an NVMe PCIe SSD based on
- IP-Maker released
NVMe FPGA IP to enable use of enterprise performance SSDs in low wattage "no
CPU" embedded systems.
In May 2017
- Rambus said it was
working with Microsoft on the design of prototype super cooled DRAM systems
to explore avenues of improvement in latency and density due to physics effects
below -180 C.
- Liqid announced
it had secured $10 million in Series A funding.
In June 2017 -
Toshiba (whose future
had been a persistent background story during the quarters leading up to this)
began sampling the world's first 64 layer QLC (x4) nand flash memory. The 768Gb
chips were the highest density nvms available.
- Micron announced
it would enter the rackmount SSD market. Micron's ambitions revolved around
SolidScale - a 2U box stuffed with NVMe SSDs and interconnected in a fabric
using software from Excelero.
In July 2017 -
Viking began shipments
of 50TB planar MLC 3.5" SAS SSDs (the highest capacity drives in that
form factor) based on a controller platform designed by rackmount SSD maker
In August 2017 - A
new SSD controller
company - Burlywood - emerged from stealth mode and won a best of show award
at the Flash Memory Summit.
In September 2017
In October 2017
- Toshiba announced the
winner of the long drawn out $18 billion
pageant to sell its memory and SSD business.
In November 2017
announced a $95 million funding round round. This proved that the
hybrid storage appliance
model for the enterprise - in which SSDs and HDDs both played a part - was still
valid as a business model. The
shortages of 2017 had demonstrated that solid state storage makers couldn't
make enough SSDs with their current production plants to sustain the needs of
the SSD market even at inflated prices.
- It was the best of times. It was the worst of times.
2017 was a year like
no other in 40 years of SSD history.
demonstrated a memory controller for the emerging Gen-Z memory fabric.
- Diablo Technologies
ceased operations and filed for bankruptcy. StorageSearch.com noted the passing
of the Memory1 (flash as DRAM tiered in an DIMM form factor concept) and said -
among other things:- "investors must have thought that if the company can't
make buckets of money at a time of
shortage and high raw DRAM prices then there would never be a better
announced the availability its new "Immortal" brand of radiation
hardened SSDs for the military and aerospace markets. The Immortal range of SSDs
included options to operate from -55ºC to 125ºC.
In February 2018
- SMART Modular
announced it had been awared a patent for technology which allowed its NVDIMM
to independently initiate a backup after an event which results in a system
freeze or "hang" without power loss (patent number
9,779,016). Active data in the Autonomous Self Refresh NVDIMM is saved and
recovered in events such as an OS crash, CPU fault, MCU fault, BIOS hang, blue
screen, or other motherboard failures.
- ioFABRIC announced it had been
awarded a patent for latency aware software. This technology is used in its
Vicinity policy engine for creating and managing data volumes based on latency
requirements - even when a volume is spread over multiple nodes and storage
- The Gen-Z Consortium announced
that the Gen-Z Core Specification 1.0 is publicly available on its website. Its
memory media independence and high bandwidth coupled with low latency enables
advanced workloads and technologies for end-to-end secure connectivity from node
level to rack scale. Following in the multidecade footsteps of Infiniband and
then recently PCIe memory fabrics I think that Gen-Z was born with with
confident expectations that because of
SCM DIMM wars
and the memoryfication of the enterprise there is a real business potential that
could become the mainstream instead of sitting in the side lines of HPC.
- Nallatech entered the in-situ SSD
market using Xilinx FPGA accelerator cores installed in NVMe compatible drives
in 2.5" and HHHL form factors.
- STT announced a
breakthrough design improvement to MRAM. STT's Precessional Spin Current (PSC)
structure - while adding only about 4nm to the height of the pMTJ deposition
stack - lengthens retention time by a factor of over 10,000 (so that 1 hour
retention becomes more than 1 year retention) while also reducing write current.
STT said this would make create new applications for MRAM in datacenter ASICs.
- Micron began sampling the industry's first SSD built on quad-level cell
(QLC) NAND technology.
- A research paper looking at workload for consumer devices in a Google
software environment showed that processing in memory could halve the
electrical power and halve the execution time by reducing data movements
compared to conventional processor architectures using similar memory.
In June 2018 -
GridGain began beta
sampling its in-memory cache as a cloud service.
In July 2018 -
Micron agreed to a
parting of the ways on future 3DXPoint development. Upto that point Micron had
seen minimal revenue from 3DXPoint whereas the potential of developing a
software ecosystem centered around this technology had strategic lock-in
attractions for Intel regardless of near term memory sales. Zsolt Kerekes -
founding editor of StorageSearch.com suggested "the memoryfication of
processors and the flattening of latency by SSD infrastructure means that
traditional complex multi level cache server processors are wasteful... and that
due to trends in memory accelerators for the
defined software market... memory and processor companies (like Intel and
Micron) will have more reasons to become competitors rather than collaborators
in strategic designs in the cloud."
In August 2018 -
Marvell began sampling
the first NVMe-oF SSD Converter Controller. aimed at a EBOF (Ethernet Bunch
of Flash) applications.
In September 2018 -
SNIA (Storage Networking
Industry Association) entered the computational storage market.
In October 2018 -
market research company
which tracked memory and
SSD price trends)
said the supercycle of DRAM price growth - which had lasted for 9 consecutive
quarters - was over. (See
cycles in the memory market for other historic examples and perspectives.)
- SMART Modular
demonstrated a 96GB Gen-Z Memory Module which was implemented in a PCIe form
factor and used bridging technology based on IntelliProp's Mamba fabric memory
|understanding SSD history?|
by Zsolt Kerekes, editor StorageSearch.com (1998 to 2018)
A you can see
above - a lot of stuff happened in SSD history. How do you make sense of that
today? Take a look at may last major article about the SSD market (written
October 2018 prior to retirement) -
ways to split SSD history into "before and after" to understand now
|About the publisher
- 27 years of enterprise buyers guides. ||
|If you could go back in
time and take with you a load of memory chips and SSDs from today
(along with compatible adapters so they could plug and play) how would that
change the world?
What if you could bring back SSDs from the future?
|what's the value
of infinitely faster RAM?
|SSD timeline - very
1999 - SSD market exceeds 10
active oems for the first time
2000 - world's first online
ads for enterprise SSDs
2001 - 3.5" SCSI flash SSDs have 14GB capacity
2002 - 1st
NAS flash SSD
2003 - terabyte SSD systems become commercially
2004 - StorageSearch.com asks - what do SSD buyers want?
- Samsung declares SSDs to be a strategic market
shipments of SAS SSD storage (server based SDS)
2006 - SSD awareness
flares into notebooks
2007 - PCIe SSD
shipments begin in enterprise
...........first designs of 3.5" SAS SSDs unveiled
...........enterprise flash IOPS can replace entry
level RAM SSDs
2008 - SSD market
reaches 100 active SSD companies
2009 - SSDs
match hard drives in capacity at 2.5" terabyte level
2010 - SSD
revenue reaches billions of dollars
...........first fizz of the SSD
2011 - SSD
software becomes useful
excitement from Fusion-io's IPO
2012 - EOL for RAM
SSD market (SSD is flash),
R/W DSP flash vendors passes 10 companies
- integrated rack level SSD technology beats the sum of the parts
...........SSD software promises to become gateway for
...........DRAM DIMMs is new
form factor for fast flash SSDs
ecosystem - negates need to invent everything by new startups
in-situ SSD processing
SSD market re-energized with 2.5" and M.2
random access memory doesn't have to be RAM
3D nand flash -may be tough enough for industrial markets
2015 - retiring
and retiering DRAM was one of the big new SSD ideas
- NVMe over Fabric, memory systems customization, etc
in legacy memory made "emerging nvms" look better
2018 - First QLC
SSDs ship. MRAM retention got 1,000x better.
first SSDs for use in Intel PCs|
|In 1982 - SemiDisk Systems (based in
Beaverton Oregon) became the first company to ship SSD accelerators for the
Intel microprocessor based PC market. |
SemiDisk's first disk emulator
cards were S-100 form factor RAM SSDs with 512 kilobytes capacity ($1,995
price at launch). Internally they used 64Kb DRAM, worked via a proprietary
interface - and were designed to work with
S-100 computers (which had Z80 8
bit CPUs inside with a 64 kilobytes addressible bus).
| Soon after, SemiDisk built
similar cards for TRS-80 Model II computers, and IBM PC's, and Epson QX-10
computers, and increased the capacity to 2 megabytes when 256Kb DRAM became
SemiDisk's founder - Jim Bell - told me (in 2014)
"the R/W throughput of these SSDs was limited by the Z-80 block transfer
function speed. (INIR, OTIR) So it was about 500 kilobytes/second speed. But
that was considered VERY fast in the early 1980's... Speed was always the main
benefit (and reason for buying)... It was mostly consumers and businesses who
Jim also said "One notable feature of these
cards is that they 'carpeted' the card with memory sockets that were literally
as close together as it was possible to push them! They used 'dipguard' type
decoupling capacitors, too."
Jim said he designed the prototype
of the SemiDisk in "about October 1980" - when he was working for
Intel in Aloha Oregon. He went on to say - "I built a device on a
wirewrap, prototype card, using 2118 5-volt 16kbit DRAMs. It had 32 sockets,
and I installed the chips 8-high. (Looked like a brick!; heavy, too,
because the chips were ceramic, specifically 'cerdip' devices.)"
|the SSD story - market
survival of the fittest?|
emerging size of
the flash SSD market as you see it today was by no means inevitable. It owes a
lot to 3 competing storage media competitors which failed to evolve fast enough
in the Darwinian jungle of the storage market in the
One of these 3 contenders is definitely on the road to extinction -
but could one of the other 2 still emerge to threaten flash SSDs?
SSD's past phantom
demons explores the latent market threats which hovered around the flash SSD
market in the past decade. They seemed real and solid enough at the time.
|| Getting a realistic
perspective of flash SSD's past demons (which seemed very threatening at the
time) may help you better judge the so-called "new" generation of nv
memory contenders - which are also discussed in the article. ...read the article|
founder writes about 20 years of DoD SSDs|
|Editor:- September 20, 2010 -
founder of Texas
Memory Systems has written a
paper (pdf) which
describes how variants of the company's newer SSDs like the
RamSan-630 have been used
recently by the US DoD and Intelligence Community.|
he describes some features of their 1st DoD SSD in 1988. The company launched
its 1st commercial enterprise SSDs in 2001 - but has continued evolving its
defense based array processing capabilities.
Later:- in December
2011 - I talked to Holly Frost - who says the SSD market is the most
exciting place to be working - about
a wide range of
subjects related to SSD design and the SSD market.
History of Solid State Memory Storage (pdf)|
|This is a guide written by
accurately sums up one of the problems which faced SSD accelerator makers from
the late 1980s to mid 1990s period until
FC SANs got established.
Here's an extract...
"Over time many other vendors have
entered, and left the business of providing solid state storage systems. Solid
state developed a reputation of being really great, if you could afford it.
This was due to an inherent flaw in the basic design theory. In short, the flaw
arises from the fact that while the storage mechanism of a solid state device is
fast, the interfaces required to connect to a host system were not fast enough
to take advantage of the performance potential.
"During this timeframe (and continuing to the present time) most
solid state systems employ slow SCSI connections. During the 1980s, 1990s and
into the 2000s, SCSI
was a slow interface and could sustain data transfer rates of only 40 to 80 MB
per second. At these speeds, host computers could not really realize the
benefits necessary to capitalize on the very high initial investment solid
state storage is expensive!"
|when did it become clear
that SSDs would be a huge market?|
|Editor:- SSDs have been listed
in the buyers guides I've published since 1992. And in 1998 - I published the
world's 1st continuously updated directory of SSD oems (the url of which is now
our main SSD news page).|
talking to people in the SSD market and in processor chip companies (as part
of my SPARC market
acceleration work) in the late 1990s and early 2000s I guessed that SSDs had
the potential to become an economic mainstream solution (instead of an
expensive niche) to counter the flattening of the peak performance growth
curve in enterprise CPUs.
But how big would that SSD market be?
When I started talking to enterprise SSD companies about my market
2003 (at that time "enterprise
SSD" was synonymous with "RAM SSD") they were initially
skeptical. Each SSD company was only seeing a small part of the market. But my
SSD pages were acting as a focus point for all vendors and most users in the
industry. It was easy for me to get an overview picture which no one else was in
a position to see.
I talked to many SSD company founders at that time -
explaining my ideas, learning more about what their technologies could do. And
they told me about customer success stories which their customers didn't want
to publicize - because it would give their customers' competitors too many
insights into how they had solved strategic business problems using SSDs.
a consensus view started to emerge from these many 1 on 1s - some SSD
companies (including some in their early years of stealth mode) adapted
their business plans around the concept of a greatly expanded SSD market
A few years later - in
2005 - I
published a new version of my SSD market penetration model which looked at
all the possible market segments for SSDs instead of just the enterprise.
That model was precipitated by the steep dive in flash memory pricing which
meant that flash SSDs would soon be 100x cheaper than just a few years
Marketers in flash SSD companies liked the new model even
better. And were happy to tell me privately how useful it was. They even
started to quote from it - although by the time the text had made it into their
sales collaterals and web sites the original attribution was mostly lost.
Where are we now?
to upcoming change in SSD than DIMM wars
| Skyera's history of
flash memory and storage video|
|Editor:- March 27, 2014 - I was a few weeks
late in discovering this talking heads SSD video... But I don't think 2 weeks
matters too much when the subject is the
history of flash memory
and storage (video) - and the main speaker is Frankie
Roohparvar, COO - Skyera - who's been
in the non volatile memory business for 30 years and has over 480 patents (in a
conversation with David Davis
a really non technical way (which even a
VC or lawyer can
understand) Frankie Roohparvar covers the ground from the earliest nvm (cavemen
making marks), lists the iconic use cases for flash based products like ipods
and cameras - which kept this market alive and innovating - and brings things
right up to date with the business thinking inside Skyera's
petabyte scale SSDs.
|If your education didn't include semiconductor
physics - and you're still struggling with imagining how
MLC differs from SLC and
eMLC, and how
controllers and all those electrons (locked in leaky cells) inter-relate
to each other - Frankie's verbal explanation will make it all fit into place.
On the business case for Skyera's approach - Frankie Roohparvar
reaffirmed something which I think the company has always been clear about - "When
90% of the system cost is flash you really need to understand the
internal workings of flash (to drive the cost down to a new lowest level)."
...listen to the video
|flash storage in notebooks
|I originally stated that - in January 2006
the first notebook maker to qualify flash SSDs*.|
I later added the
note "for use in Windows XP, Linux and Solaris notebooks."
to Robin Harris, editor StorageMojo.com
for this email note (April 19, 2006).
Omnibook 300 offered a PCMCIA flash disk as a several hundred dollar
option ($400?) back in (I think) 1993.
"I know because I bought it and used one for years. The option
had 10MB of capacity and HP packaged in a compression utility that
automatically compressed everything on the flash card, so the effective
capacity was 20MB.
"The real benefit wasn't weight, as the 300 weighed in at 2.9lbs
with or without a hard disk. The win was battery life - which went to 10
hours with the SSD from about 3-4 hours with the HDD.
"With an instant-on feature that really worked, and a decent PDA
and terminal emulation, built in Word & Excel (to which I added
Powerpoint) I had a very solid, unfussy machine that I only had to charge
every few days. Lived with it daily for 5 years until I had to give it up
because it would no longer do what I needed."
speaking the Omnibook drive wasn't an SSD, because it didn't include
it was an early example of
flash replacing hard
disk storage in a notebook style product.
|Let me introduce a new
concept in computer architecture - SSD CPU Equivalence.
The SSD accelerator market should be viewed as a replacement for part
of the server CPU market.
Not as a percentage of storage spend.
|could SSDs become a $10 Billion
|sugaring MLC for the
|When flash SSDs started to be used as
enterprise server accelerators in 2004 - competing
RAM SSD makers said
flash wasn't reliable
RAM SSDs had been used for server speedups
- and in 2004 they owned the enterprise market. (Before 2004 - flash SSDs
weren't fast enough and had mostly been used as rugged storage in the
markets - and in space
constrained civilian products such as smartphones.)
By 2007 it was
clear that the endurance
of SLC flash was more than good enough to survive in high
caches. And in the ensuing years the debate about enterprise flash SSDs shifted
to MLC - because when systems integrators put early cheap consumer grade SSDs
into arrays - guess what happened? They burned out within a few months - exactly
Since 2009 new
technologies and the combined market experience of enterprise MLC pioneers
like Fusion-io and
demonstrated that with the right management - MLC can survive in most (but
still not all) fast SSDs.
Now as we head into 1X nanometer flash
generations new technical challenges are arising and MLC SSD makers disagree
about which is the best way to implement enterprise MLC SSDs.
type of so called "enterprise MLC" is best? Can you believe the
contradictory marketing claims? Can you even understand the arguments? (Probably
And that's why marketing is going to play a bigger part in the
next round of enterprise SSD wars as SSD companies wave their wands and reveal
more about the magic inside their SSD engines to audiences who don't really
understand half of what they're being told.
|From 2007 it started to
become clear that controller architecture could play a bigger part in the SSD
pricing equation than memory - and by 2011 the activity in the market (measured
by product success) showed clearly that SSD-aware software too had the potential
to differentiate SSD products in business ways which de-emphasized the role
which memory played in characterizing SSDs (in nearly all markets). |
|an SSD view of
past, present and future boom bust cycles in the memory market|
|Reliability is more than
just MTBF... and unlike Quality - it's not free. |
|I had flagged Storage Reliability as a long term
strategic concern for the market in a trends article (2005) - in which I said
that the risks posed by uncorrectable data failures due to systemic design flaws
in storage drives "could be more serious than the Y2K bug threat - if not
dealt with in advance."
Most people didn't understand what I was talking about.
(wrongly) assumed that they could always depend on oems to design a workable
level of reliability into their storage products. And if that wasn't good enough
- then a wraparound layer of RAID supported by some type of data backup would
work well enough for their needs.
In 2010 - as we got sucked into the
SSD market bubble we
began to see more customer concerns about the poor reputation which some leading
storage oems were acquiring - due to shipping undependable and incompletely
verified SSD designs. storage
reliability (2006 to 2012)
there ever a Golden Heroic Age of the Enterprise SSD market?
the plot of the enterprise SSD story have been rewritten with different
characters to make it simpler?
musings from an
|From 2008 onwards as the
growing visibility of the expanding SSD universe moved into the start of the
first years of the enterprise SSD market bubble - the market moved into a new
phase of multiple interpretations, explanations and misinformation about the
economic cost benefits of enterprise flash.
Many SSD vendors themselves were clueless about the real economic
benefits of their products - because they didn't know enough about the user
application experience - and the diversity of user businesses and risk profiles
Astrological Age of Enterprise SSD Pricing|
|the impact of file metadata
on applications performance|
|Editor:- October 5, 2016 - The interdependency
of storage software,
physical storage and applications performance is one of the constantly recurring
themes in the SSD era. |
When you change one variable - such as the
assumption that storage has a single type of characteristic - the old west 6
shooter rotational model which underpinned all data IO requests in legacy
systems software - to a faster machine gun model - with futuristic
options verging towards the light saber and photonic data shots - thanks to
solid state storage - then data systems architects have more degrees of freedom
around which they can optimise system effectiveness.
A recent blog
from Primary Data
Applications by Offloading File Metadata Operations - gives a top level view
of why it can be useful to offload file metadata operations from storage
arrays. Among other things the blog reminds us that...
array CPUs can spend significant processing cycles keeping up with all the
metadata operations such as managing file permissions and tracking open and
close events that might otherwise be used for processing data I/O. In some
cases, this can bring both storage and client application performance to a
|data dematerialization in
|Editor:- July 27, 2016 - Some of the big SSD
ideas in recent years have been:-
One way to interpret the essence of Symbolic IO's
architecture - which was partially unveiled in
May 2016 - may
be as a coming together of the 2 concepts in the same place...|
got me thinking this way was a recent blog -
look at Symbolic IO's patents - by Robin Harris on
his site - StorageMojo.com .
Symbolic IOs founder Brian Ignomirello who saw
and liked Robin's post - said among other things on
- "yes we (do) materialize and dematerialize data." ...read
PS - In a conversation I had about the market
yesterday I noted how during the past year the SSD industry has been thinking
systems architecture as the next emerging core for innovation in the same
kind of way that the market used to buzz about apps acceleration using SSDs on
the SAN and in servers (via
PCIe cards) 8 to 10
It's because of all that previous market experience with
PCIe SSDs in servers
especially and the comparisons with other ways of getting similar results with
arrays of SAS / SATA SSDs in storage - and the feel-good confidence from having
made those difficult changes - that the data computing market is now
receptive to being more ambitious with re-engineering memory.
business incentive being that the gap between what is possible and what is
being done every day with current products is so huge and
(Huge savings for users. Multi-billion dollar new markets for vendors.)
already seen a lot of different approaches coming down the pipe in the past year
with technology announcements. But even though the implementation details are so
different - they're tackling the same problem.
With more toys in the
memory, SSD and software tool kit - there are now more permutations for
servers which exceed previous performance limits and slash away at legacy
The DNA of the semicondata market continues its
ruthless quest of doing more for less.
|a winter's tale of SSD
market influences - from industrial flash controllers to HPC flash arrays - set
against the tapestry of a single company which divided to face the tempests of
|Editor:- November 15, 2016 - Recently I spoke
President - Charles Tsai.
We talked about many changing influences in the SSD market. I thought
you might be interested to see some of the things we spoke about in a new blog
on StorageSearch.com -
tale of SSD market influences - because it will give you an idea of how
many strategic changes in the SSD market can now influence every business
decision about what new products to create - even when those changing factors
seem at first to be only loosely connected like the flash controller,
industrial SSD, SCM, software and enterprise rackmount SSD markets.
those factors entwined the flow of this SSD conversation which really
started 2 years before. ...read the
|"At the technology
level, the systems we are building through continued evolution are not advancing
fast enough to keep up with new workloads and use cases. The reality is that
the machines we have today were architected 5 years ago, and ML/DL/AI uses in
business are just coming to light,
industry missed a need."|
|From the blog -
Memory Centric Architecture by Robert Hormuth,
VP/Fellow and Server CTO - Dell
EMC (January 26, 2017)|
|Xitore envisions NVDIMM
tiered memory evolution |
|Editor:- February 7, 2017, 2017 - "Cache
based NVDIMM architectures will be the predominant interface overtaking NVMe
within the next 5-10 years in the race for performance" - is the concluding
message of a recent presentation by Doug Fink , Director of
Product Marketing - Xitore
Generation Persistent Memory Evolution - beyond the NVDIMM-N (pdf)|
Among other things Doug's slides echo a theme discussed
before - which is that
new memory media (PCM, ReRAM, 3DXpoint) will have to compete in price and
performance terms with flash based alternatives and this will slow down the
adoption of the alt nvms.
comments:- Xitore (like others in the
SCM DIMM wars
market) is working on NVDIMM form factor based solutions and in this and
they provide a useful summary of the classifications in this module
However, the wider market picture is that the retiring and
retiering DRAM story cuts
across form factors with many other permutations of feasible implementation
So - whereas the NVDIMM is a seductively convenient form
factor for systems architects to think around - the competitive market for big
memory will use anything from SSDs on a chip upto (and including) populations of
entire fast rackmount SSD boxes as part of such tiered solutions - if the
economics, scale, interface fabric and
software make the
cost, performance and time to market sums emerge in a viable zone of business
risk and doability.
ain't what it used to be