| leading the way to the
new storage frontier
|Virtium launches 32GB ULP
RAM with industrial screening|
Editor:- February 26, 2014 - Virtium today
new low profile 16GB and 32GB DDR3 RAM modules in these form factors -
ULP (Ultra-Low Profile) - 0.70" high and VLP (Very-Low Profile) - 0.738"
high - for use in dense embedded servers which need industrial temperature
operation in these form factors.
Editor's comments:- as with
- one of the differences in industrial memory is the screening process.
Virtium's memory blog
says that using 24 hour TDBI (Test During Burn-In) chambers can reduce early
failures up to 90%.
involves many complex dimensions - so too does server RAM reliability - as
these articles show...
Hybrid Memory Cube gets x2 speedup
25, 2014 - Although the market for Hybrid Memory Cube compatible
RAM has barely begun - a
new Gen2 specification was
today which doubles the fastest short-reach data performance (previously
15Gb/s) upto 30Gb/s. See also:-
A3CUBE unveils PCIe memory fabric for
PCIe SSD architectures
Editor:- February 25, 2014 -
PCIe SSDs can now
access a true PCIe connected shared memory fabric designed by A3CUBE - which exited stealth today
of their remote shared broadcast memory network -
RONNIEE Express -
which provides 700nS (nanoseconds) raw latency (4 byte message) and which
enables message throughput - via standard PCIe - which is 8x better
Editor's comments:- I spoke to the
recently - who say they intend to make this an affordable mainstream
The idea of using PCIe as a fabric to share data at low
latency and with fast throughput across a set of closely located servers
isn't a new one.
The world's leading PCIe chipmaker
PLX started educating
designers and systems architects about these possibilities
few years ago - as a way to elegantly answer a new set of scalability
problems caused by the increasing adoption of PCIe SSDs. These questions
- how do you make this expensive resource available to more servers?
the least year or so - we've seen most of the leading vendors in the enterprise
PCIe SSD market leverage some of the new features in PCIe chips - to
implement high availability SSDs with low latency.
- how do you enable a simple to implement failover mechanism - so that data
remains accessible in the event of either a server or SSD fault?
But although there
are many ways of doing this - the details are different for each vendor.
- until now - if you wanted to share data at PCIe-like latency across a bunch
of PCIe SSDs from different companies - located in different boxes - the
simplest way to do that was to bridge across ethernet or infiniband. - And even
though it has been technically possible with standard software packages - the
integration, education and support issues - compared to legacy SAN or NAS
techniques would be extremely daunting.
That's where A3CUBE comes into
the picture. Their concept is to provide a box which enables any supported PCIe
device to connect to any other - at low latency and with high throughput -
in an architecture which scales to many thousands of nodes.
heart of this is a shared broadcast memory window - of 128Mbytes - which can be
viewed simultaneously by any of the attached ports.
ever used shared remote memory in a supercomputer style of system design at
any time in the past 20 years or so - you'll know that the critical thing is how
the latency grows as you add more ports. So that was one of the questions I
Here's what I was told - "The latency is related to the
dimension of the packet for example: In a real application using a range of
64-256 bytes of messages the 3D torus latency doubled after 1,000 nodes.
With larger packets, the number of nodes to double the latency becomes grater.
But the real point is that the latency of a simple p2p in a standard 10GE is
reached after 29,000 nodes.
"A more clear example of the scalability of the system is this.
Imagine that an application experiences a max latency of 4 us with 64 nodes, now
we want to scale to 1,000 nodes the max latency that the same application
experience will became 4.9 us. 0.9 us of extra latency for 936 more nodes."
Editor again:- Those are very impressive examples - and demonstrates that the
"scalability" is inherent in the original product design.
didn't want to say publicly what the costs of the nodes and the box are at this
stage. But they answered the question a different way.
Their aim is to
price the architecture so that it works out cheaper to run than the legacy
(pre-PCIe SSD era) alternatives - and they're hoping that server oems and fast
SSD oems will find A3CUBE's way of doing this PCIe fabric scalability stuff -
is the ideal way they want to go.
There's a lot more we have to learn
- and a lot of testing to be done and software to be written - but for users
whose nightmare questions have been - how do I easily scale up to a 10,000
PCIe SSD resource - and when I've got it - how can I simplify changing
suppliers? - there's a new safety net being woven. Here are the
coming soon - an incrementally denser rackmount SSD from Nimbus
February 25, 2014 - What does a
petabyte of pure
flash enterprise SSD look like?
It depends on who makes it, the
relative speed category
of the storage itself and the user's own preferences for
schemes and associated
the densest (best) level - a petabyte of enterprise SSD can occupy as little as
2U rack space (skyEagle
The upper limit in size depends a lot on the internal
and memory type
- but typically occupies 10x to 20x more physical space than the example I've
On that theme - Nimbus Data today
that in April it will sample the next generation of its own fast
featured SSD systems - the Gemini X - which will provide an incremental 2x
capacity density improvement over the company's previous offering.
Gemini X implements 960TB (raw) unified SSD storage with 100 microseconds
latency (at under 6W / TB) in 24U of rackspace (40GB/s internal bandwidth)
with external connection via 10 dual ports which can be
40Gb Ethernet or
Like most leading vendors -
Nimbus likes to
think that its product are better than some other named competitors.
But all such comparisons are very selective.
1 in 5 strings cut at Violin
Editor:- February 20,
2014 - Violin Memory
it has downsized its workforce by 21% - compared to the peak of 4 months
ago. The company indicated that (despite this restructuring of sales,
marketing and engineering) Violin still hoped to "grow revenues by
strengthening engagement of indirect channels."
Marvell samples 5K IOPS smartphone SSD
February 18, 2014 - Marvell
it is sampling a new
5.0 controller - the 88NV1088 - which enables "SSD class"
performance (280MB/s read speed and 5K random IOPS) in a smartphone compatible
another $13 million for Primary Data
February 17, 2014 - A
in SiliconAngle.com - shared from the linkedin page of Primary Data's CMO
- Rick White -
says that Primary Data (which is still in stealth mode) has secured another
$13 million funding - bringing its total funding up to $63 million.
comments:- there's a lot of speculation about what this new company is
Primary Data's founders changed the enterprise server market
with their previous startup
Fusion-io (founded in
Before Fusion-io - server makers didn't want to talk about
After Fusion-io - no server maker could launch a new
enterprise server product line without including SSD acceleration as a standard
option. (Because Fusion-io signed up most of the key server oems as "PCIe SSD inside"
which made all server makers
hostage to the
fortunes of SSD).
When Primary Data launches its products later
this year - what will its products look like?
From the hints dropped
so far - it seems that
Primary Data will
aim to shake up the enterprise architecture world with a new platform which
leverages SSD enhanced servers as the worker ants in a software scheme which
spans everything from the local cluster to the
also:- VCs and SSDs,
enterprise SSD silos,
Diablo comments on legal barbs cast by Netlist
February 14, 2014 - Diablo's
CEO and founder -
Badalone was reported to have said in an
blog by Willem ter Harmsel today - re the lawsuit launched against
them by Netlist...
(Netlist) basically have yet to prove that:- they understand the product, and
also that they have found anything Diablo is actually infringing on. We are
just going to let this run through the (legal) process and we are confident
about the outcome." ...read
Atlantis provides more evidence of the trend towards massively
improved enterprise utilization enabled by SSD-aware software
February 11, 2014 - Atlantis
that the new "In-Memory Storage Technology" release of its storage
virtualization software - called
Atlantis ILIO USX -
can significantly increase enterprise utilization by enabling users to deploy
up to 5x more VMs on their existing storage.
USX faqs (pdf),
utilization and the SSD event horizon,
Conduant's new 3U 8TB PXIe SSD
Editor:- February 10,
2014 - Conduant
launched a 3U
single slot PXIe module which can be populated with upto 8 mSATA SSDs. The
DM-8M-3U has a PCIe Gen 2 interface which connects to the flash array via
an on-board Marvell
wanted! - VARs for Fusion's rackmounts
February 10, 2014 - Fusion-io
that its systems level (PCIe SSD inside) rackmounts will soon be widely
available from VARs in North America. Specifically these products include:-
Editor's comments:- It takes
less time to hatch a human baby than it has taken for Fusion-io to make the
transition from first talking about some of these integrated systems to
making them generally available.
Traditionally - SSD systems
companies hid the messy product creation process - and preferred to launch their
new rack babies when they were fully formed and ready to fly off the shelves.
That's in contrast to drive makers who often start revealing what their plans
are - long enough beforehand so that their customers can warm up to the idea.
Fusion's case - what we've been seeing in the past 2-3 quarters is not so much
the development of new rackmount product lines (because all the technology
components already existed before) but what we've witnessed instead is the
growth pangs and development of a new SSD systems business - jostling for
adequate space and recognition within an already confident module and software
business - with much of the thinking about the priorities being done out loud
and visible to the public gaze.
Pure Storage launches maintenance program to allay user fears
about flash array upgrade costs and obsolescenece
6, 2014 - Pure
a new business model for supporting its rackmount SSDs - the
Forever Flash program - which
the company says will provide users who have maintenance contracts better
forward visibility of costs and upgrades.
One of the clever subliminal marketing messages here - when you get into the
details - is the assumption that Pure Storage will still be in business in
future years and still offering competitive SSD systems which you'd want to
But at the same time - Pure Storage - does also tackle
head-on a key user fear which delays sales in this statement - "Flash
storage changes completely every 6-9 months, so we built the
FlashArray to allow for
incremental expansion to take advantage of fast flash density and cost
enterprise SSD box riddle game,
survivor's guide to enterprise SSDs
old software will slow new silicon in memory done by SSDs
February 5, 2014 - In a new blog -
Vistas For Persistent Memory - Tom Coughlin,
Associates reminds us that in exteremely fast SSDs - lowering the
hardware latency is just one part of the design solution.
Tom says -
"An important element in using persistent memory in the PCIe and memory
bus of computers is the creation of software programs that take advantage of the
speed and low latency of nonvolatile memory. With the increase in performance
that new interfaces allow, software built around slower storage technologies
becomes a significant issue preventing getting the full performance from a
persistent memory system."
Tom's article includes a graph which
shows the increasing proportion of the read access time taken up by system
software in successively faster hardware interface generations. ...read
Editor's comments:- living with the old
while planning for a new type of SSD-aware computer architecture is
Just how complicated that picture can be... you may
glimpse in a classic far reaching paper (about abstracting application
transactional semantics in usefully different ways when viewed from their
interactions with the flash translation layer) - called
Operations via the Flash Translation Layer (pdf) by Gary Orenstein,
reasons for fading out DRAM in the virtual memory mix
PLX is ready with 1st Gen3 compliant PCIe switches
February 4, 2014 - PLX
announced it is
the 1st PCIe switch vendor to have achieved Gen3 compliance having passed the
compliance testing procedures of PCI-SIG.
Editor's comments:- the new standard doubles the maximum
data bandwidth of PCIe SSDs to 1GB/s per lane in each direction - which enable
32GB/s total throughput for a x16 link.
SSD interface chips,
Express 3.0 Integrators List
SSD news archive -
history - since the market began
top SSD companies
top SSD articles
SSD endurance myths
and legends - now in 3D
10 key SSD
ideas which emerged and clarified in 2014
for a company called Skybox Imaging,
Inc. where I'm currently working on upgrading our storage system.
found your web site invaluable in doing market research and wanted to thank you.
- I love the mice!
Flight Software Engineer|
(email to the editor - February 28, 2014)
|Enterprise endurance - and
the SSD event horizon|
|Editor:- February 3, 2014 - I had not long
before ftped up the first draft of my new
Top SSD Companies
article last Friday afternoon - when I got an email from a reader - who
is an insider in an enterprise SSD company (and prefers to remain anonymous
here) who commented on an article I'd written a few months before - the SSD event horizon.|
said - "I had not seen that article before.... really liked it and
forwarded it to other people here. One other cause to consider for the event
horizon is this.... A real problem is that when it comes time for (system)
refresh, the customer looks at the wear life of the flash and observes
that it has plenty of available life and decides to keep it instead of
Editor's comments:- In case you haven't seen it yet either -
the SSD event
horizon article is hard to summarize - but - among the things discussed
are the predictable revenue crashes which can hit enterprise SSD vendors
from time to time - triggered when they ratchet up technical improvements in
their products which increases SSD utilization for a homogeneous user base - at
a faster rate than the sales ramp was growing before such changes.
in SSD systems endurance
wasn't one of the examples mentioned in my original article - but when viewed
from this angle - it is.
The event horizon impact on revenue is
something I had also been discussing with another company last Thursday - this
time from the perspective of
|Users will continue
choosing SSD systems which are technically very different on the inside - even
when their apps requirements (for performance and capacity and compatibility)
look superficially the same from the outside.
That's because other factors outweigh these top level headline tech
|new SSD thinking
inside the box|
|IBM Redbook places memory
channel SSDs in server context |
|Editor:- March 18, 2014 - IBM recently published a
new free 28 page ebook (aka Redbook) -
IBM eXFlash Memory-Channel Storage in Enterprise Solutions (abstract) / (pdf) -
which describes how
SSDs fit into the concept of servers relative to the other types of SSDs
already available. |
Editor's comments:- I've been writing
about this technology since the time it was being developed and have been well
briefed by the original developers - so this paper didn't have any great
surprises to me - but I think this document presents a balanced introduction
to this technology and a contextualized analysis of how it compares to the
other well established SSD acceleration options which are available for use
key takeway - in my view is table 2 - in which you can see a hierarchy of write
latencies which are approximately 5x longer in each case as you
progress up the flash SSD steps from
SSDs then PCIe SSDs
and finally SAS SSDs.
While bearing in mind that SSD data write latency is not the same
as apps performance latency (because the integration of R/W data flow
patterns with the software
plays a significant part too) and also remembering that
some products in
the market will blur the ratio of the latency boundaries for these 3
different SSD types - you can, nevertheless see why memory channel has a
distinct slot within the onboard SSD acceleration options which you need to
think about in servers.
|3D NAND flash challenges|
|Editor:- February 6, 2014 - The best article
I've yet seen about the practical implications of increasing the adoption of 3D
NAND flash is this...|
At The Table: Commercial potential and production challenges for 3D NAND memory
technology - published by Semiconductor
Manufacturing and Design.
Among the many practical
considerations discussed in this article was the question of - "how is the
semi industry preparing for the transition to 3D memory?"
issue of scalability limits and market pacing - the article reveals that
vertical scalability currently appears feasible in roadmaps upto about 100 cell
But the rate of 2D shrinks in successive 3D designs
will slow down from the recent historic average of 20% per generation to 5% -
due to the problems of registration which accumulate as you add more layers.
| "One petabyte of
enterprise SSD could replace 10 to 50 petabytes of raw HDD storage in the
enterprise - and still run all the apps faster and at lower cost."|
|meet Ken and the SSD
|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
If you're a VC should this make you anxious or happy?
hidden segments in the enterprise |
| Oh, I know SanDisk|
|Editor:- February 27, 2014 - 2 years ago - with
the acquisition of FlashSoft
- SanDisk began
to step up its efforts to reposition itself as an enterprise SSD company - an
aspiration hadn't really been articulated in earlier
(even when they had involved enterprise ready product lines).|
strategic goal of doing more in the
already been evoked in the company's investor communications a few months
before the details of how the missing IP would be sourced in detail became
transparently clear with the decision to acquire
SMART in July
blog on SanDisk's enterprise site - Jean S. Bozman,
Enterprise Solutions Manager, SanDisk - muses on how those enterprise-ward
changes are being seen through her eyes in as a relatively new employee and
also by customers - in conversations which begin with - "Oh, I know
SanDisk" (as a USB
flash drive maker) but then go on to other places.