 |
| leading the way to the
new storage frontier | |
... |
| ..... |
|
after AFAs -
what's the next box?
how fast
can your SSD run backwards?
not your
grandfather's industrial SSD market
capacitor
hold up times in 2.5" military SSDs
where are we
heading with memory intensive systems?
optimizing
CPUs for use in SSDs in the Post Modernist Era
controllernomics
and risk reward with big memory "flash as RAM"
|
|
| ... |
|
| SSD news
highlights in May 2017 |
| by
Zsolt Kerekes,
editor - StorageSearch.com |
Thinking. That's the hard part about the
SSD market. And there's a lot for you to think about in a couple of recent
stories on this SSD news page.
- MRAM? - should we still be calling this - "emerging memory"?
In
my experience a key dividing line for new technologies is when their creators
put money into setting up sales channels and advertising.
Because
although you can sell visions to
investors the real
test is selling tangible products to customers.
Everspin recently set
up an enterprise sales channel, and this month started talking about how many of
its MRAM chips are going into U.2 SSDs.
- Western Digital has paused the sale of Toshiba's memory business.
Delay
is good for WDC but bad for Toshiba and other potential buyers.
The
tensions are analyzed in recent news stories.
Would a WDC acquisition
be good or bad for the storage market?
- 3D, 3D, 3D...
Everyone is talking about their latest 3D
nand achievements (memory, SSD, array) and how they are solving the world's
problems.
You may learn a lot from a simple question posed by a
semiconductor memory company which competes with nand...
Is
Micron's 3D memory (really) cheaper than 2D?
In some ways this is
an old story which goes right back to the dawn of the
modern
SSD era (the period since when products external to the SSD ecosystem no
longer triggered extinction points for SSD roadmaps).
In past classic
articles I have discussed the differences between
SSD user value
propositions and SSD
price.
But I despite all that I anticipate you will learn a lot
about the thinking in the
semiconductor
memory industry by following the threads above.
It's a casino
in which SSDs tend to win more often than the memories in my view.
This
didn't matter when the SSD market was small - but as I wrote in 2013 -
everyone's now hostage
to the fortunes of SSD.
Even Seagate - which in
April 2017 -
tried to minimize the issue by saying that only 10% of its HDD products were
now at the risk of future substitution by SSDs.
That's an
old argument
too.
And I will return to it with a new analysis of what the
extinction threat really may be for all those future
hard drives which have
relocated themselves in the safe market havens of video surveillance and cold
storage. | | |
|
| ... |
|
 |
|
| ... |
|
| Micron dares skin in the
SSD box game |
Editor:- May 5, 2017 - Micron this week
announced
its long overdue plans for entering the
rackmount SSD market.
Micron's ambitions revolve around
SolidScale -
a 2U box stuffed with NVMe SSDs and interconnected in a fabric using software
from Excelero.
Micron is seeking early validation customers and technology partners
now and says its platform is expected to begin volume production in early 2018.
|
 |
Editor's comments:- This is a welcome
- albeit long overdue - strategic move by Micron.
Two years ago (in
June 2015) I commented on Micron's weak understanding of the enterprise SSD
systems market and I said this...
"Micron is the only company
(among the major memory and HDD companies in the same market) which hasn't
acquired any
significant enterprise SSD companies in the past 12 to 18 months. For years they
didn't understand the enterprise and they still don't have the IP or brands
needed. So they should feel strong pressure to do something. But
making a big move after years of doing nothing is risky too."
The
core of my (2015) downgrade on Micron's enterprise business was:-
It's
only possible to deliver the most efficient (and therefore competive) SSD
based products at the box level due to the accumulated incremental advantages
of big
controller SSD architecture (managing large numbers of flash devices)
compared to optimizations made at the single drive level.
And the box
level is the lowest level at which multi-layered and multi-location
intelligence can be usefully managed by internal software which is applications
and box aware and not just aware of tricks at the solo flash drive level.
Furthermore
it was clear that the enterprise market will change due to growing customer
experience and expectations leading to a
much
smaller set of systems level products supplying the whole market.
In
this predicted vein SanDisk's
Infiniflash - launched in
March 2015 -
was one of the early examples of a standard commodity SSD box being offered
by one of the big semiconductor memory or HDD makers.
Although you
can glean useful market lessons from the adoption (or not) of Infiniflash (which
was a box of SAS SSDs) and try to apply them them a prediction of the prospects
for SolidScale - I think the most useful lessons are at the tactical rather than
strategic level.
As you'd expect from any bunch of SSDs in a box -
Micron's SolidScale is currently lightweight and lacking any proprietary depth
in box scale IP.
From the technical point of view - there is nothing
that SolidScale can do which can't be done equally well and sooner by packing
similar SSDs into another branded box and using the same or similar externally
sourced software.
On the plus side - for Micron - it will learn
useful market lessons by having some skin in the SSD box game. SSD
history
shows that it takes many years before players are taken seriously by the
enterprise market (typically because of the need to integrate
HA features
and software) - but new box vendors can only get better by playing or
acquiring better players.
Customers in more technically adventurous
markets like web entities, cloud infrastructure and HPC have been the most
receptive early adopters of performance based drives (PCIe SSDs) and fast
boxes. And that's a direction to which Micron can turn to get early customer
feedback.
On the downplaying expectations side:- for
Micron...
- the PCIe fabric box is not the whole networked storage box market
- the NVMe SSD array market will itself become fragmented into many latency
zones
- the NVMe SSD array market (latency optimized storage) will not be the
whole enterprise SSD storage market.
Segmentation
challenges in the enterprise memory systems market have been complex in past
years, leaving many viable user needs unmet by satisfactory products and the
disruptive interplay of CPUs, memory and software in the
Post
Modernist SSD Era means that no current vendor business plans or roadmaps
should be regarded as being sacrosanct or endurable.
Having said all
that...
It's better for Micron to have a box business with which it
can play and learn than merely supplying the turf to other big box (stadium)
brand leaders.
Who dares wins (if they survive).
Here's a
short timeline of related articles.
- new SSD
thinking inside the box (2013) - some of the deepest thinking going on in
the SSD market is focused on rackmount SSDs, but there will continue to be
diversity in the internal architectures and detailed features due to valid
different preferences among end-users
- playing the
enterprise SSD box riddle game (2013) - for many enterprise Users - the
question - can you guess what will my next SSD box will look like? - bears a
striking resemblance to the unfairness of Gollum - when he says - what has it
got in its pocketses - Precious?
- Decloaking
hidden rackmount SSD segments in the enterprise (2014) - A contributory
cause for gaps in segmental understanding has been the continuing pace of
disruptive innovation in enterprise SSDland - which has meant there hasn't been
a stable market template for vendors to follow from one seemingly chaotic year
to the next as they encroach on new markets.
- Exiting
the Astrological Age of Enterprise SSD Pricing (2014) - vendors and users
face different sets of uncertainties and challenges - but the risk of not doing
business is bigger than the inevitable mistakes which will be made - innovative
pricing strategies can provide plausible calm assurance that everything is under
control
- we're
better than EMC - the unreal positioning of many AFA startups (2015) -
For investors in such startups - consider this. When your flash enterprise
vendor spends so much energy talking about competitive comparisons in a market
segment - which is so restricted (EMC legacy base) and about theoretical
opportunities which aren't real and which can never scale upwards (because the
big customers have spent more on EMC than your pet AFA company received in its
last funding round) - then they're wasting their share of voice in the market. .
- after
AFAs - what's next? (2017) - Every assumption which served you well about
the relative sizes of memory and storage in your enterprise world will soon
be wrong. Even "AFAs everywhere" are a transient market stopover and
no longer the final destination of the solid state storage roadmap.
| | |
|
... |
|
| BeSang poses question
about relative cost of 3D vs 2D nand |
Editor:- May 10, 2017 - I thought the SSD
industry had a long time ago answered the question - which is more important
- the memory or the SSD? - but indirectly it has been asked again in a
different way...
In a new
post
on linkedin Sang-Yun
Lee, CEO - BeSang
says...
"Micron
says 32-layer 3D NAND is about 30% cheaper than planar NAND. However, market
price of SSD shows planar NAND-based SSD is cheaper than 3D NAND based-SSD. ...Is
3D nand cheaper than planar nand?"
Editor's comments:-
My own comments to this (on linkedin) were much longer than Sang's provocative
post. But I'm hoping to learn more from what others say. ...read
BeSang's post and comment(s)
Sanity check...
From a
competitive viewpoint BeSang has been pitching their own way of getting the
memory industry to the next level in capacity at a lower price and greater
efficiency than the roadmap which Micron, Samsung etc seem to be pursuing. (And
as we've seen in SSD history and semiconductor - the best ideas for innovation
come from startups rather than encumbents. Intel itself was once a startup...
Whether BeSang's technology will get the investments needed to enable wider
adoption still remains to be seen.)
Nevertheless Sang's question is - I
think - interesting because underlying it are many difficult realities about
the bumpy transition of the memory market to a non incremental (same old
predictable spreadsheet) technology ecosystem.
When was it clear
that SSD architecture was more important than memory?
Many times as you
can see in my classic timeliney blog -
sugaring flash
for the enterprise. And every couple of years it gets said again in a
different way. Here's an example from my
2012 SSD market
transitions article.
"Knowing the generic memory type
characteristics doesn't give you enough useful info about the SSD's
characteristics and limitations any more (2012)."
PS
- Unfortuntaely you have to be logged in to linkedin to see the original post
and comments (and it has to be working). So if you're deciding if it's
worthwhile or if the original post has disapeared by the time you read this in
2027 in the
news archive in which case you already know what happened. Here is what I
commented on linkedin.
Comparing the price of planar SSDs and 3D
SSDs at the same capacity level doesn't lead to a reliable conclusion about the
competitiveness of 2D versus 3D nand.
The greater maturity and simplicity of the 2D controller will reduce
BOM cost now. But 3D (with its more expensive materials) provides a platform for
making a better SSD from the endurance view.
That's something which has user value in many market roles.
When memory production is set against a background of high book to
bill then it makes business sense for memory makers who have their own SSD
brands to bias the mix towards markets like industrial and enterprise rather
than low value markets like consumer.
On the other hand companies like Micron also know they have to keep
the consumer markets alive (even if starved) because they are the traditional
adopters of future generations of new unproven (reliability) memory.
Feed the fab can't operate if the only customers are industrial or
enterprise
That's when my finger slipped on the phone and it got
posted. I could have written more. But you get the general drift.
After
posting to linkedin and publishing the note above I told Sang-Yun Lee
privately that "The clever thing was for you to ask such a naive
(sounding) question" - because I anticipated that me and my readers will
learn much from responses.
Sang replied and posted this... "Thank
you for sharing your comment. I understand that there is no simple answer to
this question. Though, I would like to make it very simple without considering
other factors. When we just consider 'Cost-per-Bit' or '$/GB,' is 3D NAND
cheaper than planar NAND now?
It is an important question to NAND
industry because NAND goes to 72-layer already. So, technically it is
very mature. In other words, it is not easy for 3D NAND to vertically expand
from now on. So, it is right time to ask this question."
See
also:- $$D pricing! -
connects and disconnects with memory costs | | |
|
... |
|
|
|
|
| |
|
|
| .. |
 |
| .. |
|
|
NetApp acquires
Software-Defined Memory
A
report
on NoCamels.com says Network Appliance
has agreed to buy Plexistor
for $20 million.
Editor's comments:- Plexistor's claim to fame
was Software-Defined Memory - with a chip agnostic approach to
SCM DIMM wars and the
memoryfication
of the enterprise.
This acquisition will enable NetApp to play around
with options on that adoption curve in speculative system offerings without
risking too much wasted software in memory dead ends.
5 suitors for Toshiba
Editor:- May 25, 2017 - A
report
in the Japan Times says "Toshiba has told its creditor banks that it
will be difficult to select joint venture partner Western Digital as the buyer
of its chip business."
A
news
story in Nikkei names the 4 other companies still interested in buying
Toshiba's memory unit as:- KKR, SK Hynix, Broadcom and Foxconn.
See
also:- Selling
Toshiba's memory business - collected stories
the value of 5 microseconds latency for Excelero
Editor:-
May 24, 2017 - 10 microseconds is the latency advantage of Excelero's
proprietary NVMesh compared to simple NVMeoF when managing fabrics of
dispersed NVMe SSDs in a PCIe connected network. More details like this appear
in a new blog on InfoWorld -
cloud
storage architecture for the enterprise - by Yaniv Romem CTO and
Tom Leyden VP of
corporate marketing at Excelero. ...read
the article
Editor's comments:- when Excelero emerged from
stealth in March
2017 the low latency overhead of their software was a big deal - at just 5µS
compared to accessing a similar SSD in the same rack.
This is an
industry magic ballpark number which has been quoted as a worst case response
in earlier years by several pioneers in big memory architectures - including
A3CUBE and
Diablo. (Although
details may have changed or been refined since.)
SCM DIMM wars
watchers know that if you can get inside that curve for most of your worst case
latencies then (with enough memory and cache) you can run popular memory
hogging applications with better performance than using traditional DRAM in
traditionally networked servers. PCIe fabrics compete and collaborate in the
same intelligent
memory systems market. Latency lessons learned from one of these contexts
can be used to guide initial expectations (subject to verification) in the
other.
Network Appliance shipped 1EB of enterprise flash in FY 2017
Editor:-
May 24, 2017 - Network
Appliance today
announced
that in the quarter ended April 28, 2017 revenue of its all-flash arrays were
over $420 million equivalent to an annualized net revenue run rate of $1.70
billion.
The company said it had shipped over an exabyte of flash
shipped in fiscal year 2017 which is about 1,000x more than the company
shipped in 2010.
Another way of looking at it is that NetApp on its own consumed about 1% of all
the raw nand flash chip
capacity in 2016.
Symbolic IO gets ready to do business
Editor:- May
23, 2017 - Symbolic IO today
announced
the appointment of Bill
Wood as VP of Channel Sales and Strategic Alliances.
"Given
my background in the IT infrastructure market spanning over three decades,
Symbolic IO's IRIS is like no
other enterprise technology I have ever seen," said Wood. "Symbolic IO
is uniquely positioned to create a new compute category..."
new edition of the Top SSD Companies
Editor:- May 19,
2017 - The
Top SSD Companies in
Q1 2017 - was published today by StorageSearch.com.
This
is the 40th quarterly edition in this market defining
series.
We're
still seeing many signs of significant changes in the pipeline which will
sweep away and redefine the relative importance of existing SSD technologies and
product lines.
You might say - that's SSD business as usual.
And
in some ways it is. There are few rafts of stability and consistent reference in
this crazy market of ours. This series is one of them. These are the companies
which mattered most to other readers like you in the recent period. ...read the article
PS
- I accidentally had to make up a new word to describe some of the changes
we're seeing - as in the "future memoryfication of the enterprise".
Sorry
about that. If I find something better I might replace it. But for now it's the
most apt word to describe what's going on.
new patents for Corsair and Violin
Editor:- May 18,
2017 - News about the issuance of new patents in the SSD and memory market this
month include:-
- Corsair -
9,645,619
- described by its inventor Bobby Kinstle Senior
Project Manager in this way. "This one is for using tiny heat pipes to
remove heat from memory devices in really tight spaces."
- Violin
(now under new ownership) announced 2 new patents.
9,495,110
- technology behind the company's 4-way active-active RAID controllers. "Method
for managing storage memory resources for LUN allocation and related maintenance
operations within a system of high-availability distributed RAID controllers,
while maintaining coherency and high input/output speed."
9,547,588
- "Method for managing the compression and/or encoding of digital data
being stored in non-volatile memory, such as flash memory, to improve the
performance of the solid state memory cells with respect to program/erase
operations. This method may reduce power consumption and storage space usage,
while enhancing the wear lifetime characteristics of the memory media."
TrendFocus reports sequential decline in raw storage drive
capacity sold in Q1 2017
Editor:- May 16, 2017 - TrendFocus today
announced
market highlights related to storage drives in Q1 2017.
The number
of exabytes shipped in HDDs and SSDs fell sequentially by 5% but was 19%
higher than the year ago period.
TrendFocus says several factors
contributed to the decline:- including weak PC sales, cyclicality in data
center spending, and tight NAND supply affecting SSD buying patterns.
Re
delays in storage spending in data centers:- TrendFocus says...
"Some
large hyperscale/data
center companies purchased fewer HDDs and SSDs in the quarter, temporarily
slowing exabyte growth in the segment responsible for driving the highest
long-term rate of capacity demand."
Editor's comments:-
whereas higher prices of flash SSDs due to fab production capacity
constraints and lower than expected yields in new memories has undoubtedly
been a significant factor in the market - another thing to keep your eye on is
the potential for big users to get more use out of the same raw capacity by
using newer software.
That's a factor which can surprise the market
at any time as I discussed in my warning article about the
impact of the SSD
software event horizon.
Another long term trend which will depress
storage drive sales will be the refocus on
memoryfication
architectures.
I'm not going to attribute this 5% decline in
raw storage drive byte shipments to the adoption of new memory
architectures. It's probably just a correction with many causes.
But
looking ahead to the next several future years you should not be surprised to
the impacts on the storage drive market which could be much higher.
Toshiba's memory business sale - tensions in the timing
Editor:- May 15, 2017 - progress updates about the sale of Toshiba's SSD and memory
business have been part of the underlying news fabric during the first half of
2017. Tensions in the timing and conclusion of a prospective sales come from
various causes:-
- Toshiba needs the promise of a buyer as soon as possible to avert further
downgrading of the parent company due to losses in other markets.
An
additional urgency is that the memory market was in a
boom at the start of
the spin-off process - with prices for memory rising.
If the market
changes back to oversupply or if other factors reduce the premium value
placed on raw flash - before an acquisition is agreed - then the value of the
business will be less.
Delay therefore works to the advantage
of a buyer and to the disadvantage of the seller.
- The size and strategic nature of the transaction means that
geo-political
considerations and anti-trust hurdles will play big part in the decision
process.
If the potential buyer triggers a monopoly investigation in
their own operating markets then delay in the acquisition is inevitable and
could take upto a year.
- A complicating factor is that Toshiba shares considerable flash wafer fab
assets and IP with Western Digital which were part of arrangements evolved in
partnership with SanDisk -which WDC acquired in 2015.
In the current
situation where Toshiba's memory business is for sale - there could be
significant differences in self interest between WDC and any future owner of
the Toshiba memory, SSD and HDD business. In that context these
reports about WDC's growing unease can be more easily appreciated:-
- Nikkei
Asian Review (April 21) - "Western Digital remains opposed to any sale
of Toshiba's memory unit to a third party."
- FT
(May 15) - "Western Digital is taking its row with Toshiba to the
International Court of Arbitration, as the US storage technology group tries to
prevent what it claims is a breach of the terms of its joint venture with the
embattled Japanese conglomerate."
Editor's comments:- It's
clear that in the matter of who buys Toshiba's memory, SSD and HDD business
the strategic and business self interests of WDC and Toshiba (parent
company) are different.
But what would be best for the industry?
My
own view is that if WDC succeeded in acquiring all of Toshiba's memory and SSD
assets in 2017 or 2018 the effect would be to reduce the ability of the SSD
market to be as competitive internally and also (although less important) reduce
the ability of the SSD market to compete as aggressively with the hard drive
market.
Another factor is the effect of competition inside the memory
market and the effect on the emerging market for tiered big memory.
This
is a market which is working to the benefit of users because there are pressures
for flash to displace DRAM and a market climate which is nurturing the growth of
new nvms which (although small in revenue terms now) could in the long term
displace a fraction of the flash or DRAM markets.
In 5 years time
such an acquisition (a major memory supplier being acquired by one of the two
remaing hard drive companies) may not matter so much - because the SSD and
memory systems market will have moved to new levels of efficiency and systems
level utilization. But right now - I can see more negatives than positive for a
competitive SSD market if such a deal were to go ahead.
See also:-
Selling
Toshiba's memory business - collected stories
Everspin quantifies MRAM cache opportunity in AFAs
Editor:-
May 12, 2017 - Everspin Technologies
spoke about various design ins and design wins for its MRAM in enterprise flash
arrays in a conference call
re
Q1 FY17 results. Sizing the market opportunity the company said that a
typical U.2 flash SSD in a storage array can use between 5 to 9 of its devices.
And in revenue terms this would mean upto about $1,000 of MRAM revenue in a
typical rack designed by its customers.
Editor's comments:-
Everspin's first publicly disclosed design win in an AFA system was in 2014 -
Skyera - whose
buyer - WDC -
invested in Everspin in
January 2015.
Liqid gets $10 million in Series A funding for PCIe fabric
management platform
Editor:- May 10, 2017 - Liqid today
announced
it has secured $10 million in Series A funding led by Marker Hill Capital. Along
with previous seed funding from
Kingston Technology,
Phison Electronics, ABR
Capital Management and DH Capital, the company has now raised $20 million
Datacenter
architectures have seen significant advances with the advent of the cloud era
but have remained shackled by the motherboard/chassis paradigm, which locks
resource allocation to the point of purchase. Through innovations in low-latency
fabrics and intelligent software, the Liqid CI Platform addresses this painful
and costly limitation by leveraging pools of compute, networking, data storage,
and graphics processing residing on the PCIe fabric, managed by Liqid CI
software, to deliver transformative results.
Editor's comments:- Among other things the SSD IP amassed by
Liqid includes an NVMe successor to the racktop PCIe fabric switch box which
PLX was talking
about in 2014 - according to
a
blog about Liqid's composable infrastructure by Robin Harris (StorageMojo).
See also:-
SSD software,
PCIe SSDs,
VCs in SSDs and storage
from 2000 to 2017
V&G releases Kylin compatible rugged VPX SSDs
Editor:-
May 10, 2017 - If rugged 6U SSD cards for military radar and flight recorder
applications or fast data acquisition is your thing (as it used to be one of my
things about 30 years ago) then you may be interested in 2 new products
released by . |
 6U SRIO VPX (pdf)
8TB MLC / 4TB SLC with 2 lanes of RapidIO host interface, approx 3GB/s
sequential R/W, fast secure erase. |
 6U PCIe VPX (pdf)
16TB MLC / 8TB SLC with PCIe 3 x 8 host interface, upto 5GB/s read, 4GB/s
write, and random R/W IOPS (4K) 200K / 160K respectively. |
Other OSes supported include VxWorks, Solaris,
Linux etc.
Editor's comments:- for a few years now I've had interesting
conversations with Limuel
Yap , VP - Global Business Operations - V&G about the various
aspects of the custom
SSD market and especially the ways in which SSD companies which have a
strong
focus
on the military market create product lines which can be just as interesting as
the standard products for publicly known applications that we read so much about
on the web.
The ingenuity and problem solving of these companies
largely remains unknown for good reasons.
Following up on the new
products above Limuel told me - "Developing these products will separate us
from other SSD suppliers that only have the standard SSD. This marks down our
capabilities to offer all kinds of memory storage most specially for military
and data recorders users."
Netlist secures funding for patent dispute with SK Hynix
Editor:-
May 4, 2017 - Netlist
announced
it has obtained outside investment to finance the legal fees and costs of its
legal action against SK
Hynix related to patents for RDIMM and LRDIMM memory technologies.
Editor's comments:- patent disputes are expensive but
apparently Netlist has secured what is in effect a no-win, no-fee deal.
An
initial determination is expected in October 2017.
A later
news report
(May 30, 2017) suggests that Netlist may also pursue a claim related to
NVDIMM technologies.
Everspin revs up its enterprise sales channel
Editor:-
May 4, 2017 - Everspin
today announced the appointment
of Annie Flaig as
Senior VP of Worldwide Sales.
Prior to Everspin, Flaig was VP of
Worldwide Enterprise Storage Solutions Sales at
SanDisk. Before joining
SanDisk, Flaig was VP for Americas Sales & Channel Marketing at
HGST driving revenue
and market share growth of HDD and SSDs in the Americas region. And before
joining HGST in 2012, Flaig spent nearly 20 years with
AMD where she served as
the Corporate VP of Worldwide Commercial and Enterprise Sales.
Editor's
comments:- this is an interesting evidential marker in the timeline of
market readiness and shippability of Everspin's MRAM as it changes phase from
being an "emerging" memory technology to one which has definitely
emerged.
more readers aren't scared of mice
Editor:- May 2,
2017 - The monthly readership of StorageSearch.com
increased by over 100% in April compared to a year ago. Thanks.
SMART files for IPO
Editor:- May 1, 2017 -
the parent company of SMART
Modular Technologies today announced that it has
filed
a registration statement on Form S-1 with the U.S. Securities and Exchange
Commission ("SEC") relating to a proposed initial public offering of
its common stock.
SMART has applied to list its common stock on the
NASDAQ Global Market under the ticker symbol "SGH." Barclays and
Deutsche Bank Securities are acting as lead book-running managers for the
offering; Jefferies and Stifel are acting as book-running managers; and Needham
& Company and Roth Capital Partners are acting as co-managers.
| |
| . |
| . |
 |
| . |
|
| |
| .. |
|
|
.. |
|
|
|
.. |
|
| Are we
there yet? |
| After more than 20 years of writing guides to
the SSD and memory systems market I admit in a new blog on
StorageSearch.com -
Are
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
construction". ...read
the article | | |
|
.. |
|
| 15 years ago in SSD news |
In May 2002 AMD announced it was sampling
64Mb Mirrorbit flash with 100k P/E endurance and 20 years retention at 125
degrees centigrade.
The next time we reported on Mirrorbit was in
October 2004 when AMD's NOR flash memory was the smallest technology in
production at 110nm.
AMD's flash business was spun off as Spansion
in 2003 (as a joint venture with Fujitsu).
Spansion went into chapter
11 in March 2009 and emerged again in May 2010.
Then in December 2014
Spansion (and the Mirrorbit flash technology) was acquired by Cypress
Semiconductor.
Cypress
's Mirrorbit (in 2017) is a 256Mb is a high reliability, temperature
tolerant flash memory with operating temperatures from -40C to +125C.
see
also:- charge-trapping
MirrorBit NOR flash | | |
|
.. |
|
Memory in China – from
Chips to Cloud
Emerging Memory Architectures
Lunch |
| Editor:- May 5, 2017 - Two interesting things
and one very essential thing (above) which I noticed today in the agenda of
the forthcoming Memory+
Conference which will take place June 1, 2017 in Shanghai. | | |
|
.. |
|
|
|
.. |
|
|
|
.. |
|
|
|
.. |
|
| after AFAs?
- the next box |
Throughout the
history of
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.
 A
new blog on StorageSearch.com
- cloud
adapted memory systems - asks (among other things) if this will always be
true.
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
article | | |
|
.. |
|
|
|
.
. |
|
|
|
.
.
.
. |
|
| 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. |
| what's
RAM really? - RAM in an SSD context | | |
|
.
.
.
. |
|
|
|
.
.
.
. |
|
| I think it's not too strong
to say that the enterprise PCIe SSD market (as we once knew it) has exploded and
fragmented into many different directions. |
| what's changed in enterprise
PCIe SSD? | | |
|
|
| . |
.
.
.
. |
|
| The same memory block may
have different ECC codes wrapped around it at different times in its operating
life - depending how healthy it looks. And different ECC codes may be used
within the same memory chip at the same time. |
| Adaptive flash
care management & DSP IP in SSDs | | |
|
. |
|
|
|
| |
