 |
|
| leading the way to the
new storage frontier |
..... | |
|
|
| |
| ..... |
|
For more pages like this see
storage history
month by month |
| .. |
new SSD racks announced
(and unannounced) by Violin
Editor:- June 24, 2014 - Violin Memory today
launched
a new rackmount SSD -
Concerto
7000 - which can provide upto 280TB of raw fast
high availability
enterprise flash (using Violin's standard
VIMM
modules) in 18U.
The key thing about the 7000 is that it's a software
services rich storage box -which enables many types of replication functions
to be easily scheduled and run concurrently (if so desired) between various
7000 nodes in dispersed geographic locations to ensure data survivability.
Violin
also offers an upgrade kit which enables existing users of its legacy 6000
storage arrays to interoperate with these services.
Editor's comments:-
June 25, 2014 - yesterday in this space I promised I would give you a bigger
story about Violin.
I said - it wouldn't be about this new product -
the Concerto 7000 - which I regard as an incremental and predictable gap filler
- but instead about an interview I conducted with the company last week
driven by the need to understand their business strategy better and to assess
how they aimed to fix this paradox...
Each time Violin (or any
enterprise SSD vendor) adds new market optimized software and features into a
new product line - it makes the new product more attractive to some users - but
much less competitive for most users.
Reconciling that marketing
paradox with Violin's aspiration to significantly grow revenue led me to pose
some very direct and detailed
segmentation
questions to the company. And I got some very interesting answers.
here
it is
In early to mid June more than the usual number of
investors had been asking me questions about Violin which I mostly answered
by referring them to scattered fragments of content which talked about their
themes - on the pages of StorageSearch.com
With another quarterly
update of the Top SSD
Companies coming soon - I was already thinking it might be a good idea to
freshen my profile page for Violin - when I got contacted by a
cheerleader from
the company asking if I wanted to talk again to Eric Herzog
who was planning another round of editor / analyst briefings about a new
software related product announcement (see above).
I said - "thanks
I'll be happy to talk (again) to Eric. But my main interest is not Violin's
new software (I talk to over 100 pure play
enterprise software
companies). However - if Eric is happy to divide his time between talking
about the new product launch – and my random questions about Violin
(prompted by many hours of discussions with readers about Violin in recent
days) – then let's make a date."
And that's how this meeting
got set up.
As things turned out - we used nearly all the allocated
time talking about my segmentation issues - and with less than 3 minutes
remaining at the end of that I said to Eric - I get the idea (about the
software in the Concerto 7000). I'll read up about it.
BTW - I didn't
send Eric any advance notice of my detailed questions.
That's partly
because I prefer thinking on the spot and we've spoken about enterprise
SSD segmentation before so I know he thinks fast.
I thought it
might be a nice break for him to get away from the standard presentation
treadmill.
To be honest I wasn't really sure how much of our
conversation I would eventually be able to write about either - because I was
touching on some serious business issues.
And what's good for me to
know - when I'm talking in my self appointed pastoral role as the virtual
best friend of every SSD company - isn't the same as what's good for the world
at large to know - especially competitors.
Anyway - the people I talk
to regularly know that although I'm a bit strange (who else would be crazy
enough to spend so much time thinking about this market) I can be relied on
to keep their secrets.
On my part I was hoping to get clarification
on 2 issues.
- Business limiting factors in the server based SSD acceleration market.
As
Violin had made it clear it wasn't go to play in the PCIe SSD market - and the
company's VIMMs don't suit a memory channel SSD form factor either- the
company's only publicly known stakeout in this market is the recently launched
WFA.
While that's an interesting experiment - which will appeal to some
kinds of users who like Windows and who like the idea of having a single
source integrated SSD enhanced apps server - and while it gives Violin a
definitely different personality to other vendors competing in this space - I
could see 2 problems which would limit the scale of this business.
1 -
the fact that Violin provides the servers.
Most big users will
prefer to standardize on servers from their preferred server suppliers.
2
- the market limiting effect of the Windows SQL environment.
Because
there are other OS's and other apps.
- Business limiting factors in the SSD storage array market.
In
particular I was wondering how Violin was going to deal with customers of the
type who liked its earlier array products - specifically because they didn't
come with a lot of software strings attached.
My concern was - that
that Violin's new product roadmap looked like it was going down the road of
increasing product specialization and increasing the software bundle.
Now
despite what some naive market reports about the SSD array market may tell you
- more functionality equates to a worse matching product for those customers who
prefer to use a different preferred set of software.
Because for
customers in that group - who don't want it - all it does by being in the
bundle is slow things down, add to their costs and support and integration
headaches. So I opened my conversation with Eric by saying I had been
speaking to a lot of people about his company recently and one of the main
things I wanted to talk about was what Violin was going to do about the
apparent segmentation problem - which is - the more you specialize a product for
one group of customers - the less attractive you make it for everyone else.
Because
I couldn't see the solution to this question coming from a continuing sequence
of product variations like the last 2 products.
And to make sure we
were on the same page I said - the outward sign of real marketing strength -
is being able to say with confidence which segments you definitely want to
walk away from.
He said straight away - we already did that when we
sold the PCIe SSD business.
Which showed we were on the same
wavelength.
So I quickly gave Eric a summary of the concerns which I
have already outlined to you above. And this is a summary of what he said.
On
the server side - he said he has seen those exact same issues.
He
named some customers (which I won't divulge here) who liked the look of the
WFA and went as far as evaluating it. And when they did that they liked the
performance they got.
(He reminded me BTW that the initial published
performance numbers for the WFA came from Microsoft - and not from Violin.)
But sometimes he conceded - that Windows is just a small part of a
customer's OS mix. If they're a big organization it's too cumbersome to start
supporting additional types of systems components in their infrastructure.
And
- on the SSD storage array side - he readily agreed with my point that there
are many customers who prefer using their own software which they already have -
rather than introducing a new bundle of software like that in the 7000. (And he
went on to give me a lot of examples of such use cases which Violin has seen
in its customer base.)
By then he'd already told me about how Violin
was already satisfying those needs with an unannounced new product.
Eric
said - I am not going to walk away from that kind of business or those kinds of
customers.
In a way - when Eric first told me the details - I was
surprised. But I wasn't shocked.
Because I had been thinking that the
solution might lie in the direction of a stripped down - no frills - software
light SSD array.
But I had anticipated that might be a future product.
Like a Strad 2000...
Can I talk about this? - I asked.
Yes
- but don't name the customers or the other details about the software
environments we talked about - said Eric.
So this is where I can reveal
the details of Violin's new product - which satisfies many of the problems I
mentioned above - which arise from having too much specialist software in an SSD
array.
It's the
6000
series.
Although - not exactly.
Eric said - we can keep
the 6000 competitive and up to date - because most of the modules (power
supplies, fans, memory) are interchangeable with the 7000.
In a way the
solution is obvious. But in another way it's not.
I said to Eric - the
usual thing in this industry is that when a company brings out a new product
which looks a bit like the product they had before - but which does more - then
everyone assumes that the old product is obsolete.
He agreed.
So
maybe there's a case to made for relaunching the 6000.
The reviewers
- who simply look at the bullet points and count how many different software
features it's got - will hate it.
But the customers - who just want a
fast reliable and predictable SSD box - and who don't want to be told what
software they have to run with it - will be delighted.
BTW - I
deliberately didn't ask Eric to break out numbers for WFA sales.
And I
won't try to estimate them by comparing them with similar things to guess a size
of market.
I said to Eric in this conversation - you don't need
analysts or market reports to tell you if the WFA experiment is working. You're
going to know that soon enough from your sales people.
So if it's still
in the product line next year - that will be a clue to us all.
See
also:- playing
the enterprise SSD box riddle game,
Decloaking
hidden segments in the enterprise for rackmount SSDs,
What makes
this enterprise SSD different?,
after AFAs -
what's the next box?
Avago agrees to acquire PLX
Editor:- June 23, 2014 -
PLX Technology
today announced
that it has agreed to be acquired by Avago Technologies for
approximately $309 million.
"The core PLX PCIe silicon business
fits very well with the Avago business model and broadens Avago's portfolio
serving the enterprise storage and networking end markets," stated Hock Tan,
President and CEO of Avago. "Following the closing of the transaction, we
are excited to welcome the PLX team to Avago, and we are committed to continue
to invest in the PLX PCI Express platform."
Editor's
comments:- to get a taste of the big ideas to come in PLX's PCIe fabric
vision see - an
SSD conversation with PLX
what's different about Tegile?
Editor:- June 23, 2014
- Tegile is a
relative newcomer to the pages of StorageSearch.com
(first SSD news citation
in February 2012)
and - frankly - I didn't think too much about the company to begin with -
but I do now.
You can see the gory details of how and why I recently
changed my mind in the new article -
an SSD
conversation with Tegile.
will there be a bidding war for Fusion-io?
Editor:-
June 23, 2014 - I've been asked by several readers if I think there might be
companies thinking about launching a bidding war for Fusion-io.
13 months ago -
where there was similar speculation - I came to the conclusion that the kind of
companies which would be a good fit (at that time) to acquire FIO - flash
memory makers and enterprise software companies - either didn't have the
knowledge about the enterprise to make such a move likely (in the case of
memory companies) or (in the case of enterprise software heavyweights) had
other, more pressing concerns.
In the current state of the market,
however, I would be amazed if several companies weren't actively considering
such an option. Although considering it - is a long way from deciding it's
a good enough idea to action.
And I would add to my earlier list
of company types likely to have such an interest (in considering their
options) would be some big server oems. (Although I think they would get
mostly get their needs satisfied more cheaply by simply letting the current
deal with
SanDisk go ahead.)
Whether
any of that analysis will translate into actions - we'll just have to wait and
see.
It depends how the calculations work out - not only about the
technology possibilities - but compatibility with business cultures and
customers too.
...Later:- a few days after writing the above -
an
article in Wall Street Journal reported that Fusion-io had held acquisition
talks with 11 companies in the period between October 2013 and May 2014.
IBM is #1 in rackmount SSD revenue
Editor:- June
16, 2014 - IBM
announced
today that a recently published market report by IDC identified IBM as the
#1 company (ranked by revenue in 2013) for rackmount SSDs with 25% market
share.
Editor's comments:- this would only be a surprise if
you had not read my January article -
Who's who in SSD?
It's IBM Jim - but not as we know it - which was cut and pasted from this
SSD news page you're reading now. .
The IBM-Jim article included
IBM's FlashSystems shipment numbers and the revelation that for 3 quarters IBM
shipped everything they could make.
See also:-
rackmount SSDs,
market research
SanDisk to buy Fusion-io...
What will SanDisk really get
from Fusion-io?
Editor:- June 16, 2014 - SanDisk today
announced
a definitive agreement to acquire Fusion-io in
all-cash transaction valued at approximately $1.1 billion.
Editor's
comments:- the result of combining the product lines from SanDisk and
Fusion-io will be an enterprise SSD product offering which is unmatched in the
industry in a broad range of enterprise SSD product categories including:-
from
Fusion
from
SanDisk:-
The acquisition will enable several new things which would
not have been technically possible or profitably viable from either company
on its own including:-
- Fusion-io's software is scalable downwards into more market segments
(such as consumer and commodity enterprise array
components such as
2.5" PCIe SSDs)
which were not attractive for FIO to pursue on it own.
But for a
flash memory company like SanDisk the ability to create
controllerless SSDs -
would enable entirely new product types to be built. - SSDs built on such
technology (without internal microcontroller offload processors) would be the
cheapest in class - compared to all traditional SSDs with internal CPUs.
- Fusion-io's rackmount storage virtualization software will open up new
markets for SanDisk's SATA SSDs.
For example in the
no-frills
embedded rackmount segment SanDisk would be able to offer customers a
rackfull of SATA SSDs as a basic integration component - thereby simplifying
integration and support issues for customers who need large quantities of
vanilla storage racks.
And Fusion-io's hybrid array product line - is
a ready made platform for spawning a new high end storage array - which
could use low cost flash SSDs instead of hard drives. At the high end
business level - for SanDisk - the ability to monetize flash at the rackmount
systems level (which is the most efficient way to convert raw flash chips into
usable enterprise SSDs) is a rational next step from its acquisition last year
of the world leading adaptive controller technology from
SMART.
All
memory makers are operating under self imposed constraints on investing new
capacity for flash wafer
starts - while they evaluate what comes next.
So one way to raise
the revenue ceiling from the same raw flash is to own
big controller
architecture with high
utilization
effectiveness (like that from Fusion-io) which multiplies upwards how
many petabytes of virtual usable flash can be delivered to satisfy users
compared to other (more) wasteful designs and business channels.
That
consideration - the ability to get more enterprise petabytes out from the same
raw flash chips in - by shipping it through better architecture - is a more
significant business factor in the flash memory market today than the ability to
do another cell geometry shrink - or adding a few more layers of toppings on
the 3D nand pizza.
summary
What will SanDisk really
get from Fusion-io? - more usable flash petabytes out from the same raw wafer
starts - due to better architecture.
and here's some more
For
those interested in the emerging
memory channel
SSD market - I've written an article - which speculates what could happen
to this class of flash DIMM SSDs in various hypothetical contexts - such as -
whether or not SanDisk also acquires
Diablo.
See:-
MCS versus PCIe
SSDs (another slight return)
scalability as a market penetration tool
Editor:-
June 16, 2014 - New SSD vendors become part of the furniture in user
environments by a variety of doorways.
Hammering at the door with bold superiority claims is one way - but
another is for SSD companies to edge their way in by incremental advances
which begin as offers to patch up the old furniture and help it last longer.
A new
blog
(Letting it all scale out) by Andy Warfield, CTO Coho Data provides a
revealing narrative about how Coho's sales people have been edging their way
into new customer sites between major purchasing cycles - and explains why
scalable systems can not only reduce risk - for some customers - but also lower
the cost - by deferring the need to buy everything sooner than it's needed - and
taking advantage of the trend that each terabyte of flash you can defer
buying today will cost you much less in the future.
"One thing
that has been remarkable about Coho's customers is that they have frequently
been buying our product off-cycle" says Andy Warfield in an educational
framework article which will help you understand one of the routes by which
this type of technology is penetrating the strongholds of legacy technology.
...read
the article
WARNING! - CONSUMER SSD
contents liable to change without notice
Editor:- June 13,
2014 - 6 years ago when I published the article -
Can you trust flash SSD
specifications & benchmarks? - which began with this sentence - "One
of the things I've noticed is that the published specs of flash SSDs change a
lot - from the time products are first announced, then when they're being
sampled, and later again when they are in volume production" - I was
alluding to the unintended negative performance consequences of firmware
upgrades rather than the deliberate substitution of an entirely different
SSD controller
introduced in an SSD product family soon after its launch.
But it
seems that the risk of deliberate component substitutions by the original
branded SSD maker (rather than merely the supply chain risk of counterfeits
by persons unknown) is another uncertainty which readers in the
consumer SSD market
may now have to contend with - according to a new article -
SSD
shadiness - vendors caught switching to cheaper components after good reviews
written by Joel
Hruska in an article for ExtremeTech.
After
naming and shaming specific SSD offenders with various accused offences -
Joel goes on to warn other vendors who might be thinking of doing something
similar - "If reviewers can't trust that the performance they see is
the performance end-users will receive, they'll never recommend your components."
...read
the article
Editor's comments:- A reader in the
consumer market contacted me recently because he thought I didn't give enough
warnings about the potentially unreliability of consumer SSDs.
I
was shocked by that view and sent him a deluge of links to cautionary
articles - including those below - in which the extracts shown here are the
original unedited quotes from each article.
- Why can consumers
expect to see more flaky flash SSDs? - SSDs aimed at the consumer market are
designed to deliver basic functionality at the lowest price. That means the
designers (originally due to ignorance – but nowadays with foreknowledge)
have to decide what shortcuts they can take in the production process and what
design factors they can leave out to reduce the price - compared to a reliable
industrial / military / enterprise grade SSD.
- the consumer SSDs
guide - How good are consumer SSDs? You won't find anyone more enthusiastic
about solid state storage than me. But - here's an important sanity check. Even
the very best consumer SSDs available today are vastly inferior in performance
and reliability to the best SSDs in the enterprise and industrial markets.
- overview
of the notebook SSD market - In the first 5 years of its history (2006-2010)
the notebook SSD market was a disappointment to SSD evangelists like me -
because integration with PCs was so bad, and most of the SSDs were too slow or
had too little capacity to be useful...
And there are more - but I
won't list them all here.
The danger for newcomers to many SSD
market is they read 1 or 2 articles about
endurance
and think that's all there is to know about SSDs.
If the SSD world
was as simple as that I would have abandoned this topic 10 years ago instead of
writing all those other
SSD
articles.
say Hi! to highly available CacheIO
Editor:- June
10, 2014 - CacheIO
today
announced results of a
benchmark which is
described by their collaborator Orange
Silicon Valley (a telco) as - "One of the top tpm benchmark results
accelerating low cost iSCSI
SATA storage."
CacheIO says that the 2 million tpm benchmark on
CacheIO accelerated commodity servers and storage shows that users can
deploy its flash cache to accelerate their database performance without
replacing or disrupting their existing servers and storage.
Editor's
comments:- The only reason I mention this otherwise me-too sounding
benchmark is because although I've known about CacheIO and what they've been
doing with various organizations in the broadcast and telco markets for over a
year - I didn't list them on StorageSearch.com before.
That was
partly because they didn't want me to name the customers they were working with
at that time - but also because with
SSD caching companies
becoming almost as numerous as tv channels on a satellite dish - I wanted to
wait and see if they would be worth a repeat viewing. (And now I think they
they are.)
PS - I asked Bang Chang,
CEO of CacheIO if he had a white paper which talked more about the company's
cache architecture and philosophy. He sent me this -
CacheIO High
Availability Deployment (pdf) - from which I've extracted these quotes...
- re network cache appliances - "At CacheIO we believe that network
cache appliance is the best storage architecture to decouple performance from
capacity and achieve the best of both worlds.
Once deployed as a "bump
in the wire" performance accelerator, our network cache appliance can also
deliver additional value added services... Compared to server-side Flash cache,
our network cache appliance is a shared resource that is more scalable, more
reliable, supports clustered applications, and most importantly allows
customers, especially cloud service providers, to monetize performance by
dynamically allocating resources based on changing SLAs."
- re operational transparency - "Implementing CacheIO network appliance
requires no change to existing applications, servers, or storage. CacheIO can be
slotted in, turned on to accelerate applications, and turned off if necessary,
often without needing to stop the applications."
I found it
interesting to see that in addition to conventional connections (SAN and
InfiniBand) their HA
paper also mentions emerging PCIe fabric.
PCIe SSDs versus memory channel SSDs are these really different
markets?
Editor:- June 10, 2014 - Throughout
storage history
we've seen many claims that a new upstart storage technology will replace
another.
It doesn't always work out to be as simple as that.
In
the enterprise SSD market of today -
PCIe SSDs are the safe
standard legacy server acceleration SSD technology whereas
memory channel
SSDs - aka UlltraDIMM, TeraDIMM, MCS etc are the upstart unproven (in the
mainstream market) wannabies.
Each of these 2 SSD types can do some
very similar things - such as make server apps run faster - but they aren't
the same.
It's reasonable to ask - are there any hard technical rules
where you can say for your new projects either:-
- I definitely don't need to know about the new thing?
OR
- maybe I should take another look at it?
And can you make
a safe (for now) yes or no decision without getting mired in contentious
product by product comparison claims and counter claims - which will suck away
hours or days of your life?
Yes you can. See my blog -
memory channel SSDs
versus PCIe SSDs (slight return) - which was on the home page of
StorageSearch.com throughout May 2014 and which now has a new permalink
location and some updates. ...read the article
PLDA enters the PCIe SSD controller market with 6.6 Gb/s NVMe
controller
Editor:- June 9, 2014 - PLDA today
anounced
immediate availability of a new
SSD controller aimed
at the PCIe SSD market.
PLDA's
XpressRICH3
IP is a high performance, low latency, highly-configurable PCI Express 3.0
soft IP supporting endpoint, root port, switch and bridge configurations. The
certified design has been engineered for both ASIC/SoC and FPGA implementations.
Editor's
comments:- PLDA has written a product description -
Conquering the
challenges of PCIe with NVMe in order to deliver highly competitive Enterprise
PCIe SSDs (pdf) - which contains an overview of the design. Here are some
extracts.
- power saving features? - PLDA implements L1 PM Sub-states, reducing power
consumption without affecting performance.
- performance? - 270K write IOPS, 12µS write latency write (4KB) using
the FPGA-based reference design.
Here are some questions I put to the
company.
- is there a minimum order quantity or cost associated with licensing your
IP?
No minimum quantity.
- What is your deliverable?
An Encrypted Source code file that can
be downloaded or sent on CD
- How does a customer evaluate your IP?
We provide 1 month Free Evaluation IP
new PCIe SSDs from Fusion-io
Editor:- June 5, 2014 -
Fusion-io
today
launched
2 new gen 2.0 x8 standard height, half length PCIe SSD product lines - which
they call the Atomic Series
- which are available in 2 different price and application categories.
- PX600
series - upto 5.2TB usable capacity, with 92µS / 15µS R/W
latency, 2.7GB/s / 2.1GB /s R/W throughput and 375K write IOPS - designed for
high end servers - with 5 years warranty.
- SX300
- upto 6.4TB usable capacity - with similar latency to PX600 - but lower
throughput (2.6GB/s / 1.2GB/s R/W throughput) and 16% lower power consumption -
designed and priced for scale-out architectures and hyperscale environments -
with 3 years warranty.
Fusion-io says that more than 7,000 customers in
over 80 countries use its products to reduce latency and speed the flow of
data-driven applications.
Editor's comments:- Unlike some other
companies Fusion-io doesn't change its core hardware products very often. The
top level message from the company - which is supported by
quotes
from many of their big partner / customers who love the products - seems
to be - it's like what you had before but better.
These new modules are
reported to use 19nm to 21nm nand flash. Other competitors - particularly in
the SAS SSD market and
memory channel
SSD market have been shipping products with that kind of flash geometries
in enterprise SSDs for over a year. So why the delay?
I
wrote about this
anticipated lag in Fusion-io's memory adoption and the likely causes,
effects and remedies 2 years ago.
In line with my reported
expectations at the time - this memory adoption lag hasn't prevented FIO from
being able to compete in the market while using older, more expensive and
reliable raw flash memory. That's because FIO has
efficiency and
utilization gain
effects in its software - which from the business point of view have
offset those low level cost disadvantages.
I'm guessing these flash
memory adoption delays were due partly to the original deficit in IP (which was
the same for nearly all SSD companies a few years ago) but also because
Fusion-io's software rich architecture - means it has to be much more
careful about integrating new complexities in its flash management - so that
the new functions don't negatively impact all the other things which were
working well before. Fusion-io can't just simply buy a new
controller (or a
controller company) - because the software is the controller.
What
these new Atomic memory generation products do for Fusion-io is reduce the
recurring raw cost of PCIe SSDs (in this performance class) and facilitate
higher capacities in the same footprint.
From the business point of
view this will have 2 effects.
- increase the profitability of each PCIe SSD sale
- increase the density and competitiveness of Fusion-io's high end fast
rackmount SSDs
Overall - you can regard these products either as a "catching
up" excercise - if you're a semiconductor memory competitor which competes
with Fusion-io - or as a "getting ahead (again)" exercise - if
you're a partner or an investor.
 Fusion-io
has also changed their logo to add the phrase "At the speed of NOW"
which I think does efficiently communicate a credible idea of what the
company has always been about.
You may recall that a year ago -
WhipTail was implying
a similar sense of urgency with its message "data at the speed of life".
If you're interested in
speed related
messages and metaphors in SSD brands - I've got an article which includes
many more such examples.
Fusion-io has also produced a
video to accompany the
new "At the speed of NOW" branding.
Now - I'm not a fan of
SSD videos - and
this doesn't convert me into one.
It has good production values - as
you'd expect from the company which made video feasible on the world's biggest
web sites - but the best thing I can say about the message content in this
video is that if you mute the sound and switch off your video screen you
won't miss anything of any value.
If you've been thinking about going to the Flash Memory Summit -
you can now do something about it
Editor:- June 5, 2014 -
Registration has
now opened for the Flash Memory
Summit - which takes place in August in Santa Clara.
Topics
will include:- NVMe | Enterprise SSDs | Flash in Data Centers | Flash on the
Memory Bus | Enterprise Storage | SSD Technology | Controllers | Memory Channel
Architecture | Data Recovery | PCIe SSDs | Application Performance | Flash and
the Internet of Things | New Non-Volatile Memory Technologies
and
a lot more besides.
SMI supports 16nm nand with new SATA SSD controller
Editor:-
June 4, 2014 - Silicon
Motion today
announced
support for Micron's
16nm 128Gb nand flash
with its
SM2246EN
SATA 3 SSD controller which enables 67K / 65K R/W IOPS (4KB) while also
enabling low power designs (2mW DevSleep mode and 21mW Slumber mode).
is there a market for I'M Intelligent Memory inside SSDs?
Editor:-
June 4, 2014 - Are there applications in the SSD market for DRAM chips which
integrate ECC correction inside the RAM chip - and which plug into standard
JEDEC sockets?
That was the question put to me this afternoon by Thorsten
Wronski - whose company MEMPHIS
Electronic AG distributes I'M Intelligent Memory in
Europe.
Thorsten told me he's had a good reaction from the SSD
companies he's spoken to - which is why he phoned.
But in a long
conversation about the economics and architectures of end to end
error correction
in SSDs and the different
ratios of RAM cache
to flash in SSDs - I told him that my initial reaction was he should look
at embedded applications - which depend on the
reliability of a
single SSD - rather than enterprise systems in which the economics analysis for
arrays point to a system wide solution rather than a point product fix.
The
interesting thing is he said he's done tests on the new I'M memory as drop in
replacements for unprotected memory designs- in which he accelerated the likely
incidence of error events by increasing the interval between refreshes and
raising the temperature.
Here's what he said.
"We
assembled a standard 1GB unbuffered DIMM with 8 chips of 1Gbit ECC DRAM. Then we
put this into a test board and ran RSTPro (a very strong memory test software).
No error found.
Next we put the whole board into a temperature chamber
at 95°C, which normally requires the refresh rate to be doubled (32mS
instead of 64mS). No error found.
Finally we wrote a software to change the refresh-register of the CPU
on the board, so we were able to set higher values. The highest possible was
750mS, so the DRAM did almost not get any more refreshes. Still it continued
working in RSTPro without a single error for 24 hours.
We tried the same with Samsung and Hynix modules, but none of them
came even close to those results. Most failed at refresh-rates of 150 to 200 mS,
which is not bad indeed. Many more tests will follow."
Editor's comments:- the reason I mention this - is because
adapting the refresh rate was one of the things mentioned in my recent blog -
Are you ready to
rethink RAM?
However - most of the leading SSDs in
industrial markets
don't have RAM caches for other reasons (to reduce the physical space, power
consumption, hold-up time, or because don't need the performance). So I told
Thorsten I don't see an industry wide demand inside SSDs. But some of you
might already have thought of applications.
See also:-
I'M ECC DRAM product
brief (pdf)
Unigen enters M.2 SSD market
Editor:- June 3, 2014 - Unigen today
announced SLC or MLC
NAND flash based
M.2 SATA III SSDs in
densities ranging from 16GB- 256GB.
These small form factor flash
modules are used for high reliability storage of critical code and data, in
space-constrained embedded, industrial, gaming and networking applications.
They are available in both commercial and
industrial
temperature ranges. | |
| . |
| SSD
history, SSD
market research,
more SSD articles
and blogs |
| . |
 |
| . |
 | |
| .. |
 |
| Megabyte loves reading news | |
|
. |
|
 |
|
. |
|
PCIe SSDs
SSD history
an SSD
conversation with PLX
Top SSD Companies -
2014 Q2
meet
Ken and the SSD event horizon
how fast can your SSD
run backwards?
|
|
. |
|
| So you'd think that when it
comes to the person in purchasing - who's looking for standard SSD products to
go into industrial or embedded systems - which aren't pushing the state of the
art when it comes to form factors (some like CF have been around forever) and
whose needs include the type of SSD interfaces which haven't been seen much in
SSD news in the last 10 years - it should be simple enough. |
| no such thing as a simple
standard legacy industrial SSD | | |
|
. |
|
 |
|
. |
|
| Decloaking hidden segments
in the enterprise |
Editor:- StorageSearch.com recently
published a new article -
Decloaking
hidden segments in the enterprise for rackmount SSDs
Some of the
world's leading SSD marketers have confided in me they know from
their own customer anecdotes that there are many segments for enterprise
flash arrays which aren't listed or even hinted at in standard models of
the enterprise market.
Many of these missing market segments don't
even have names.
Hey - that means SSD-world is like a map of the
US before Lewis and Clark.
If you're a
VC should this make
you anxious or happy?
If you're a user - maybe that's why no one is
delighting you in the way you think you deserve.
That's what led me to
write my new article.
See
also:- rackmount
SSDs, SSD silos,
market research | | |
|
.
.
.
. |
|
 |
|
.
.
.
. |
|
stories from SSD news in
June months gone by
June 2013 -
Samsung launches M.2
form factor PCIe SSD
June 2012 -
Seagate says it will
use DensBits DSP
technology
June
2011 - NVSL unveils PCIe SSD using
phase-change memory
June 2010 -
Anobit samples memory
signal processing flash SSDs
June 2009 -
Fusion-io unveils
consumer priced PCIe SSDs
June 2008 - Fast
boot RAM SSDs get patent protection
June 2007 -
SanDisk launches 64GB
notebook SSDs
June 2000 -
M-Systems packs 2.6GB
into 2.5" SCSI SSD |
|
.
.
.
. |
|
|
|
.
.
.
.
.
.
.
. |
|
| new SSD articles coming
soon |
Editor:- June 23, 2014 - Here are the outlines
of some new articles which will appear on StorageSearch.com in July.
- The credibility of endurance claims.
A reader told me -
I've got a bunch of different (named) suppliers offering me
industrial SSDs
which I know have exactly the same controllers inside and the same memory chips.
But the TBW (total
terabytes written) numbers they quote me vary wildly - 5 to 1!
In
my application the mission cost of device failures is huge. Is there a reason
for this disparity in rated
endurance?
Who can I believe?
- A useful list of military SSD companies.
Although you can
already find details for over 50 military SSD companies on StorageSearch.com -
it takes a lot of work using site search.
The new directory which will
be linked from the military
SSD news page - will include a list of suggested companies you should
look at if you need SSDs which have been designed for the special needs of
defense related embedded applications.
- An SSD conversation with Charles Tsai, Senior
Director - InnoDisk.
I
had been wanting to learn more about the internal thinking in InnoDisk for some
time.
We talked about new ways to design enterprise flash arrays and
about new ways to make industrial SSDs better which borrow concepts from
cloud architecture.
- SSD market consolidation?
That phrase has come up in 2
or 3 conversations each day in the last month.
Apart from the wishful
thinking aspect (of those wanting to be
acquired and
those doing the acquiring)
- is there going to be any real lasting difference - after a few more months
of this activity?
- a new directory for PCIe fabric
I've already written a
lot about this subject. But it hasn't had a home. There are going to be a lot
of new developments in this subject.
So next week I'm going to
datamine all my past articles about PCIe fabric and create a new dedicated news
and articles directory for PCIe fabric in an SSD context to make it easier to
find. | | |
|
.
.
.
.
.
.
.
. |
|
 |
|
.
.
.
.
.
.
.
. |
|
| |
