| leading the way to the
new storage frontier||...|
what's the state of DWPD?
simple list of military SSD companies
concerning SSD endurance
should we set
higher expectations for memory systems?
is remanence in
persistent memory a new security risk?
doesn't just provide capacity..." |
in his recent linkedin blog -
Storage First, Not Servers April 15, 2016.|
Editor's comments:- I
told Jason I agree with him about the relative importance and nature of storage
not being commonly enough appreciated.
Like mass and energy in physics
- server CPUs and SSD storage head counts are interchangeable in black box
observations of apps and data.
This is a market paradigm with market
disrupting effects which I called attention to in
2003 and is what I
nowadays refer to as the "SSD-CPU equivalence" driver of SSD market
More recently (in
list of user
value propositions for adopting SSDcentric infrasture grew to 6 core ideas
with the addition of replacing swathes of enterprise DRAM with cheaper, higher
density flash and other (long
time emerging) alt nvms.
wars phenomenon in some aspects looks like a new DRAM-flash equivalence
proposition for latency filtered tiers within complex DRAM installations.
While this "market surge" aspect is new - its technology
is rooted in classical
ideas about virtualizing RAM with mixed memory and storage types with
the new market twist simply being the credibility of semiconductor memory
roadmaps having sufficiently distinct cost/density characteristics combined
with the confidence of existing massive SSD installed infrastructures to make
it worthwhile supporting multiple memory chip types with investments in
safest time to perform garbage collection?|
March 16, 2016 - while there's often no single correct answer to SSD design
questions (see the
SSD design heresies
for examples) a recent blog -
SSD Controller Design Compromises the Best NAND Memory by Cactus Technologies
reveals an interesting insight into the design philosophy of its embedded SSD
lines. Among other things:- |
"On-the-fly Garbage Collection many
flash products do garbage collection during IDLE time to maximize performance
when the device is active; however, this increase the likelihood of write abort
data corruption as the host system may remove power during IDLE time while the
device is still doing program/erase operations internally. Cactus Technologies
industrial grade products performs on-the-fly garbage collection; while this
will reduce performance slightly, it ensures that the host can safely power down
the device during IDLE time."
Editor's comments:- While
SSD designs still have to mitigate against
sudden power loss (which like most
- the Cactus products do) it's interesting for system designers to know that
there can be multi-layered approaches to power protection designed into some
SSDs. (Instead of single shot silver bullets.)
Actually the main point
of the Cactus blog is something else entirely - the difference in correctable
ECC in different geometries of SLC. Nice graphs if you're interested.
isn't a fan of pSLC for industrial uses saying - "There is not
substantial data to back up pSLC as any more reliable than using the full
capacity of the MLC component with 2x the cells to perform wear leveling."
|10 years after... more word
on QLC SSDs|
|Editor:- March 15, 2016 - Toshiba's long
anticipated first generation of QLC (x4)
nand flash is expected to
of 500 PE cycles according to a presentation reported recently by Robin Harris in a
Editor's comments:- Although QLC
technology hit the SSD news pages
10 years ago with
nand technology from M-Systems
winning an innovation award at the 2006
Flash Memory Summit this did
not result in commercially viable products anywhere near as soon as had been
In a 2007 blog -
happened to x4? by Mark
Savolainen we learned that M-Systems had already been working secretly
on x4 for 5 years before those optimistic (pre-)announcements.
the benefit of hindsight the technology which could be demonstrated to work
in small arrays (and large cell geometries) in 2006 - needed a mind bogglingly
complicated anount of controller
R/W complexity and ECC support to enable its operation across a competitively
useful chip capacity built on mass producible processes.
The x4 data
integrity problem still looked problematic in 2010 when I wrote about the
difficulties in the context of stealth mode
of the roadmap to truly solving this problem lay in a different technical
approach to flash controller management -
flash care management & DSP IP in SSDs - which became a
hot topic for
the SSD industry in
2012 in the
context of enhancing the reliability and speed of MLC and TLC.
another reason for the >10 years delay in seeing QLC SSDs in the enterprise
will have been the market's slowly emerging readiness for such an idea - whose
friendly reception needed to be prepared in advance by the growing
supported comfort zones of DWPD
and confidence within
user segments that a latency
bands approach to solid state storage could be a sustainable
ships applications aware FT caching|
|Editor:- March 8, 2016 - PrimaryIO (which
its name from CacheBox in
the general availability of its Application Performance Acceleration V1.0
(software) for VMware vSphere 6. |
APA aggregates server-based flash storage across vSphere clusters as a
cluster-wide resource enabling all nodes in the cluster to leverage the flash
caching benefits even though a subset may already have flash deployed.
application awareness, PrimaryIO APA caches critical, latency-sensitive
application IOs in order to boost overall application performance while enabling
optimal utilization of data center server and networking resources.
APA supports write-around and write-back caching with full
in face of node failures since writes to cache are replicated to up to 2
Editor's comments:- in a
brief (pdf) about their technology - PrimaryIO describes how they use
application awareness to intercept data request streams based on its "relative
value and ability to accelerate workload performance." PrimaryIO says this
is more efficient in
its use of flash than traditional approaches and can get good
with a smaller amount of installed SSD capacity than other methods which don't
discriminate so accurately.
on the unspoilt shoreline of AFA |
a nostalgic hybrid roadtrip memoir
|Editor:- March 11, 2016 - AFAs account for 30%
business - says the company in a recent blog
back to the future which jibes at competitor Nimble's implied tardy
upmarket conversion from the
humble hybrid appliance
roots which they both share thematically if not in other ways.|
implies that more years of experience with "all flash" and its own
claim to having paddled around this market beachhead for 4 years already
make it a better dining companion or supplier for customers to depend on.
you think about it, however, longevity in flash storage box marketing
isn't the surest of clues to flash DNA blue blooded staunchness.
it were so then Violin -
which has been in the solid state storage box market nearly 9 years - would be
seated closer to the head of the AFA Camelot roundtable.
maybe - in an alternative universe - they think they still are.
flash for the enterprise (2004 to 2016)
|new funding for endurance stretching NVMdurance|
March 29 , 2016 - NVMdurance
it has completed a $2.5 million Series A round of financing. Existing investors
New Venture Partners, ACT Venture Capital, Enterprise Ireland and NDRC have
invested bringing total funding to $2.77 million.
financing builds on an exceptional year from NVMdurance which saw its first
customer announcement with
Altera (now part of
Intel)" said Steve Socolof
of New Venture Partners LLC. "The NVMdurance software increases the number
of program-erase cycles in Altera's FPGA-based storage reference design by up to
7x times compared to existing NAND flash implementations."
comments:- NVMdurance says the power behind its endurance stretching IP is
the use of offline machine learning software that automatically learns the
optimal parameter settings for the NAND device.
See also:- VCs in SSDs,
razzle dazzling flash
NVMdurance's machine learning to manual tuning
2 ASIC roles for PCIe based BiTMICRO SSD controllers
March 25 , 2016 - 5
years ago when BiTMICRO
unveiled an earlier generation of its high performance enterprise SSD
controller architecture - it was clear that their preference was for a chipset
which included 2 different types of functionality.
This kind of
thinking wasn't unique at that time - as I'd seen similar things in
designs before but (unlike BiTMICRO) those other designs were captive and not
offered as COTS SSD
How many controller chips do you really need for a
a new blog
today BiTMICRO explains why its current generation of controllers continues
using a 2 ASIC architecture with one acting as a flash array extender and the
other as the main PCIe host interface controller.
Among other things
the blog says "To increase flash channel bandwidth and capacity, more flash
channel expander chips can be instantiated and connected to the main controller."
As noted in the SSD
design heresies - SSD vendors often have different implementation
architecture approaches which compete in similar application slots. When
evaluating different types of offerings it can be useful to ask yourself - which
direction is my own design likely to stretch in future? (Towards more
performance? lower cost? bigger scale? adjacent application role? etc.)
BiTMICRO's blog clarifies where they see their strengths in the market. ...read the
Microsemi's secure SSD business to be acquired by Mercury
March 23 , 2016 - Mercury Systems
has agreed to acquire the
secure SSD business of Microsemi as part
of a $300 million deal which includes several other business units focused
on the defense
electronics market which altogether employ approximately 275 people based at
facilities in Phoenix, Ariz., Camarillo, Calif., San Jose, Calif., and West
2000 to 2016
Renice announces pSLC military SSD
22 , 2016 - Renice
the imminent availability of a 2TB 2.5" SATA MLC SSD family for the
military market. The
R-SATA uses the company's
controller. Interesting aspects of this announcment are:-
- Renice says the drive can be used in pSLC mode - which halves the
capacity but "achieves nearer SLC performance and reliability".
Editor's comments:- Renice has a completely
different view about the efficacy of pSLC in this type of SSD - than
Cactus which dismissed
the notion that it was worthwhile in a recent blog about memory geometries and
ECC which you can see on the left of this page.
- the new SSD uses rugged SATA connectors which are work more reliably in
In this context I
think the key differences in SSD design are:-
versus classical flash management and also
regular RAM flash cache ratio.
As we know from SSD history - the
data integrity and reliability you measure is just as dependent on the
controller architecture as the intrinsic flash. So once you steer a course for
your product line down these permutation avenues (due to past experience or
convenient access to you own related IP) then the consequences you get - in
terms of product envelope - are not necessarily reproducible by competitors who
chose a different road.
On the rugged SATA connector issue:- the first
time I recall seeing that being mentioned in a rugged military SSD news story
was April 9, 2012.
what does Infinidat think about the SSD market?
March 22 , 2016 - Views differ
about many things in the SSD market and one of the fundamental bones of
contention has always been - the pace of the (decade long predicted)
race towards the
solid state storage data center and how long (7 years already) there will
remain a viable market for
hybrid storage appliances
(hybrids which include hard
drives - as opposed to hybrids which consist of
multiple latency adapted
That prompted me to ask Steve Kenniston, VP
Product Marketing at
(which sells multi-petabyte hybrid appliances) for his suggestions for the
series on StorageSearch.com
You can see what Steve recommends you read today on the
home page, and if you like these
links and want to easily find them again later - his SSD bookmark
suggestions are permalinked
Avere ranked #1 in Google's cloud partner search list
March 16 , 2016 - How well does Avere Systems (and
its virtual edge filer) work as a
Google's cloud services?
Apparently very well - as Avere today
it had been named "Google Cloud Platform Technology Partner of the Year"
enterprise SSD petabytes doubled in China in 2015
March 15, 2016 - Despite all the
(doing more with less) we've been seeing due to
slashing away at the waste in utilization within
virtualization stacks - the enterprise SSD market is still growing.
Insights says that Enterprise SSD petabytes doubled YoY in the China
market in 2015. While at the same time - shipments of all types of SSDs grew 3x
faster in the China market than the overall worldwide market.
comments:- Greg was coy about giving me exact numbers when I asked - which
is why you got ratios instead in the story above - but you will be able to
find raw numbers in his new report -
China's SSD Market - which will be published next month. Whoops - I forgot
to ask the price. Most of Forward Insights'
past SSD reports
have been in the region from around $5K to $10K.
storage market research,
this way to the
closes $55 million funding round
Editor:- March 15, 2016 - Pivot3 today
it has closed a $55 million equity and bank financing round funded by
Argonaut Private Equity and S3 Ventures.
Among other things Pivot3 will
use the funding to integrate the NexGen QoS capabilities with Pivot3's leading
hyper-converged, highly usable capacity, fault tolerant solutions.
PCIe 4.0 chips milestone
Editor:- March 14, 2016 -
Signs of onwards and upwards progress towards future
PCIe SSD speeds
emerged today in an
have demonstrated electrical interoperability for PCIe 4.0 with "robust
signal integrity" (BER below 10-15) at 16Gbps with 4 lanes running
revenue depleted Violin says it will keep on trekking
March 11, 2016 - No one looking ahead to the revenue etc
yesterday by Violin
Memory could reasonably have expected to hear uplifting results and so
I doubt if many were surprised by the tone of the announcement. Revenue for the
quarter ended January 31, 2016 was $10.9 million - 47% lower than the year
a related webcast Violin said that it had embarked on several strategic
relationships with vendors which had been approached as part of the (now
abandoned) process of finding a buyer for the company as its primary way of
Violin says it is still hopeful that
evaluations of its new systems and software which are now in progress by
over 40 significant customers could result in future sales which bring the
company towards a sustainable customer funded operating position.
Violin also said that despite many industry announcements from other
directions it remains convinced that it still has technology leadership in its
newest product lines.
Editor's comments:- the competitive
challenge for all rackmount
SSD companies (not just Violin) is how convincingly (to their customers not
to themselves) they can
themselves in the roadmap towards
enterprise storage box consolidation in which a much smaller set of
standard platforms will account for nearly all the storage boxes which are
Violin's current puny revenue places it at a severe
competitive disadvantage compared to those other competitors which have their
own proprietary big
controller SSD architecture (which were developed in small companies but
which are now owned by IBM, EMC, WDC, SanDisk etc).
On the other hand
no AFA vendor - which doesn't use proprietary hardware - can feel relaxed just
because it's assembled from standard COTS drives either. Because in the
case of such vendors as Pure, Tegile, Kaminario, HP etc their challenge is
to win the battles of dominance in software platforms or place (marketing
strength in channels, brand, services etc).
Having said that (and
despite its past waste on vanity marketing) the huge development cost (just
ask Violin's past investors how much) of Violin's fast, fault tolerant storage
boxes does still buy them a place in conversations which begin with - if we
don't want to risk being held to ransom by those other big guys - is there
another way of doing this reliable flash array stuff today? (Even though we
know that everything
is changing again.)
HP offers up front guarantees on usable capacity in flash hybrids
March 10, 2016 - Taking the guesswork out of the costs of flash array
was one of the new pricing trends I wrote about 2 years ago in my article -
the Astrological Age of Enterprise SSD Pricing.
HP has aligned itself with
this trend in a
updates to its hybrid array family -
Among other things HP says - "The new HPE
Thinner Guarantee program offers a free, up-front workload assessment and a
written assurance of as much as 75% capacity savings that removes guesswork in
migrating from legacy storage onto all-flash HPE 3PAR arrays. The program,
which now includes savings from 3PAR Thin Deduplication, uses specialized, big
data assessment tools to determine the amount of capacity reduction that a
customer can expect to see by migrating to a 3PAR all-flash array. The program
uses these results to offer an individualized, written guarantee of customer
capacity savings. With this program, HPE offers the strongest assurance of
capacity reduction on the market, backed by a written contract."
Seagate promises faster PCIe SSDs
8, 2016 - Seagate
it will ship a 16 lane NVMe PCIe SSD with 10GB/S throughput in the summer. No
further details are available at this time.
in 2015 the fastest production PCIe SSDs were probably the
NVRAM from Microsemi
and the MX6300
2016 the fastest
motherboard SSDs will be
SSDs rather than PCIe SSDs - which potentially will be able to deliver
twice the performance of 16 lane PCIe SSDs - but which are more
capacity and fault
Pure says - revenue grew because customers like us and
BTW we expect to get more business in the cloud
Editor:- March 3,
2016 - Pure
it has more than doubled its revenue for the quarter ended January 31, 2016
compared to the year ago period. Revenue was $150 million for the recent
quarter and $440 million for the full year.
In a related earnings
Pure indicated that one dimension of its revenue growth came from repeat
business within the customer base saying - "For each initial $1 purchase,
our top 25 cohort purchased greater than $12 more of Pure Storage within their
first 18 months."
Strongly hinting that its systems remained
competitively priced Pure said that a quarter of its business (upto that time)
had come cloud
a question about how much future cloud market business might be
accessible to external systems suppliers like Pure compared to custom "do-it-yourselfers
like Amazon and Google who build their own storage" Pure's CEO Scott Dietzen relayed
thinking from IDC that
by 2019, 80% of that cloud market would remain non-captive and
accessible to vendors like Pure.
Editor's comments:- The
bigger problem for all such companies (not just Pure) in the 2019 timeframe is
not the scale of non-captive TAM but how they reposition themselves within a
future rackmount SSD
market which will be dominated by a
smaller number of distinct standardized platform types than today.
there will always be
for proprietary systems - Pure and others will have to decide - are they the
ones setting the new commodity standards? - or will they have to realign their
offerings to work with the standards set by others?
Apple and FBI case demonstrate difficulties of SSD data recovery
March 3, 2016 - If anyone still had doubts about how difficult it is to recover
data from an encrypted SSD in the absence of a universal back-door key - the
proposition has been lent weight by the recent story rippling around the
world's news media about the FBI's efforts to force Apple to assist in
unlocking iphones. In the unlikely event you don't know what I'm talking about
here to see summaries of the unfolding story.
multiple uses and many of them originated as part of intelligence and law
enforcement data gathering activities.
Defeating data recoverability is
a primary objective of
destruction design techniques used in many
|What happened before?
|Compared to earlier phases
in the SSD market more is changing... No wonder no one's got a clear picture.
It's a mess.|
|Your mission - should you
decide to accept it - is to compile a simple list of military SSD
manufacturers. That sounds easy - I thought. How hard can that be?|
|How hard was it? |
|Why can't SSD's most ardent
believers agree on a single shared vision for the future of solid state
|the SSD Heresies |
and the virtual memory mix slider
| Editor:- March 2, 2016 - in a new blog on
loving reasons for fading out DRAM in the virtual memory slider mix - I
cast an eye on the latency specific defects in DRAM system behavior which
are among the many technology enablers of the emerging tiered memory /
flash as RAM market. |
We've been accustomed to think of DRAM as the
simple predictable latency memory (compared to
flash). But server
motherboard memory system latency hasn't improved for over 10 years. Memory
systems got bigger and bandwidth got faster but worst case latencies can
sometimes be worse than they used to be - due to interference effects caused
by complex data queuing patterns.
If you haven't noticed these
problems - congratulations!
It means you might not notice (or care)
when the virtual memory slider moves in the cheaper direction towards
memories like flash. ...read the
|Who's got all the answers
to help understand how all the changes in the SSD market are coming together?
The answer is - no one and everyone and you too.