|
|
| "a new guide for those
seeking consumer SSDs" |
| ...headline from:-
SSD news | | |
| .. |
| Pure
Storage says what you can do with those HDD arrays |
Editor:- May 16, 2012 - Pure Storage
today published a new video
on YouTube which pokes fun at the idea of hanging onto
hard drive arrays and
suggests what you can do with them.
The 142 second video packs a lot
of humor into its tour of why their way of doing
dedupe with
flash is cheaper and better. And it includes
animals too.
The
company also
unveiled
a new generation of fast-enough
(100K write IOPS)
HA/FT SSD arrays
today - with upto 100TB compressed capacity - which are clustered around
InfiniBand. |
 |
I'm not great fan of SSD
videos - because they mostly waste time - but this one will be added to my
favorites list later
today - because it's amusing and speaks for the SSD industry. ...watch the video | | | |
| .... |
| What makes
this enterprise SSD different? |
just 4 things really... (well - maybe 6)
ending
the infinite SSD article loops!
by Zsolt Kerekes,
editor -March 30, 2012
If you've been involved in one of those
projects recently trying to decide which SSD supplier to choose for your
strategic, mission critical, server future - then you've probably read more
articles about this subject and pored over more benchmarks than you expected
when you began.
It seems like the more you read - the more you have to
know.
A lot of readers have told me they spend days reading what I've
written about SSDs and sometimes it seems like they're going around in circles
because just when they have satisfied their needs to understand one strand which
surrounds this topic they discover another piece of information which opens
up another new loop they have to master - or flatly contradicts something
which they thought they had previously understood only a few loops before.
Is
there a conspiracy of SSD bloggers and vendors to churn out thousands
of SSD articles to obscure the handful of simple SSD sub atomic particles
which really explain everything?
In one way that's true.
Reason
being - the enterprise SSD world is still evolving.
As in particle
phsyics we do some experiments based on the shape of things we understand with
the current level of technology - and then discover that there are some other
little things out there which weren't so obvious before.
So if we
create an SSD model which says that all CPUs and software only know about hard
drives - you get one type of optimal SSD solution....
But then a few
years later if some CPUs and software know about SSDs - maybe that changes how
they behave - and as the SSDs get cheaper and faster and smarter (and
everywhere) it's possible to do market experiments which wouldn't have been
feasible before.
an end to SSD article loops
If you're
tired of going round infinite SSD article loops I can offer you a small number
of indivisible SSD architecture principles - which - if you understand them -
will make it simpler for you to predict and model most of the important behavior
and quirky performance and reliability characterirstics of almost any enterprise
SSD that you will read about.
All SSD personalities are dominated by
just a handful of design parameters. If you know these SSD architecture quarks
/ quirks you can go a long way to understanding and predicting SSD
system behavior. Here's my little list of articles.
That's it. That's all you need to know.
Nearly...
Obviously I haven't said anything about the many
different types
of flash memory which can be used inside such SSDs.
Nor anything
much about
SSD reliability
engineering. These are problems for your vendor's SSD designers to worry about
and not for you. There are many different ways to achieve similar product goals
using vastly different starting points in the reliability of their memory parts
(depending how clever the SSD oems are and their business models).
And
I haven't said anything about
SSD prices or
interfaces or form
factors - because you're only going to look at SSDs you can afford,
connect to and which fit in your box. Those aren't the differences which start
the midnight SSD article loops spinning.
There are only 2 more
articles which I'll add to my list above to make it complete.
- the SSD software changes everything!
- is still in the category of "in the queue to do" rather than "FIFOed
out of the SSD content spaghetti extruder" - but you get the general idea
of what it will be about.
|
 |
Please don't complain to me
if I just started you on another one of those endless reading loops. I tried to
make it as simple as I could. Honestly. | | | |
| . |
|
|
| . |
| the top 14 SSD
companies in May? |
Editor:-May 15, 2012 - I started publishing a
quarterly list of the
top 10 SSD companies (based on search volume) over 5 years ago.
Why?
- well because I knew the market would get so big that nobody (apart from me)
would have the time to follow up hundreds of (future) wannabe SSD companies -
and so priority SSD shortlists would become essential when researching
suppliers, partners, competitors and investments.
And financial or
shipment data is either not available, or is too backward looking to be
useful at predicting future winners at the steep spiky stages of high growth
in chaotic truly disruptive markets.
FYI the top 14 SSD companies
that StorageSearch.com readers followed up in the first 2 weeks in May
were:-
1 - Fusion-io 2
- Violin Memory 3 -
STEC 4 -
OCZ 5 -
LSI/SandForce 6 -
DensBits 7 -
Texas Memory Systems 8
- WhipTail 9 -
Virident Systems 10
-
Kove 11 -
SanDisk 12 -
SMART 13 -
BiTMICRO 14 -
RunCore
A week
isn't a long time in the SSD market. But it's long enough to detect big changes
(if they coming). A month is more reliable. And a quarter is better still. Much
longer than than - and you might as well wait for the bean counters. (Although
the true message - as seen in financial reports lags the advance signals from
search volume predictions by many quarters.)
If you compare the the
top SSD company lists quarter by quarter - it seems at first as though
nothing much is changing. But if you compare over a longer period you do see
significant changes. For example some of the companies in the list above
weren't even marketing SSDs 5 years ago.
The search volume methodology
is very fast at picking up future long term winners - sometimes in the same
quarter that they emerge from stealth mode - even if it takes years for
financial data to later confirm these successes.
One established
pattern seems to be that
SSD controller companies
and SSD software
companies which have appeared in the list - get acquired by bigger companies
who want to enter
the SSD market.
A new pattern that we're starting to see
- is the importance of companies having their own SSD IP.
What about
the future of the SSD market?
Lots of web sites talk about where the
SSD market is going. Some are more reliable than others. A few years ago I
compiled a list of SSD
market research companies which I regard as the most experienced.
The
SSD market is growing fast, and solid state storage will be one of the 2
biggest items in the enterprise computing hardware budget by the end of this
decade. |
 |
Funny to think that even 3
years ago - SSDs didn't appear in most IT budgets at all. | | | |
| . |
| |
.... |
| "In May 2001
- Winchester Systems introduced a product called - FlashSSD - as an
option in its OpenRAID enterprise storage SAN product line. This non-volatile
solid-state disk was for the typical 1% to 5% of an application's "hot
files" that account for 50% or more of all disk requests. FlashSSD
delivered a sustained and constant 12,000 IOPS and 40MB/s data throughput. The
company said it could speed up disk based applications by 2x to 5x...." |
| ...from:-
SSD market
history | | |
.. |
 |
.. |
| how fast can your
SSD run backwards? |
SSDs are complex devices and there's a
lot of mysterious behavior which isn't fully revealed by
benchmarks and
vendor's product datasheets and whitepapers. Underlying all the important
aspects of SSD behavior are
asymmetries
which arise from the intrinsic technologies and architecture inside the SSD.
Which symmetries are most important in an SSD?
That
depends on your application. But knowing that these symmetries exist, what they
are, and judging how your selected SSD compares will give you new insights
into SSD
performance,
cost and
reliability.
There's
no such thing as - the perfect SSD - existing in the market today - but
the SSD symmetry list helps you to understand where any SSD in any memory
technology stands relative to the ideal. |
|
| | |
| . |
| what do
enterprise SSD users want?.................. |
by Zsolt Kerekes,
editor - March 2012
And what do SSD companies have to do in
the next few years to make it easier for wannabe SSD enterprise customers to
buy more of their products and be happy?
You'd think that someone
should know all the answers by now.
But they don't.
I talk to
many thought leaders in the enterprise SSD market - and even when they have a
clear idea of what they're doing and where their own roadmaps are going - they
don't have spare time to figure out questions like - how are all these
different parts of the future SSD jigsaw going to join up in the user
enterprise?
If users pay the invoice and buy repeat systems - that's
as good as it gets. The investors
are happy. Why worry about SSD climate change?
Many SSD gurus have said
to me recently - while catching their breath for the next sprint - it's a
crazy market.
Succesful enterprise SSD businesses are working hard and
racing fast.
The technologies which satisfied one SSD product
generation's needs rarely sustain their reliability and competitiveness edge
for more than a few years before needing to be reinvented.
But
success in SSD generation X doesn't guarantee success in generation Y.
Looking
back on the past 15 years or so of the enterprise SSD market you could say that
SSD marketers had it easy.
As long as each new product was faster,
denser, cheaper and
more reliable than
the one before - and came attached with the right interfaces - their job was
mostly done - because that satisfied the needs of the market.
Today's
SSD market is much more complicated.
Here's a simple example of top
level market market fragmentation.
Suppose you're designing a new
rackmount SSD...
In
the old days they were all fast and all were compatible with
FC SANs. Some were a shade
faster than others. Others a tad cheaper. But within the market there wasn't
the vast spectrum of capability you see today.
To compete viably
today rackmount SSDs have to be design optimized to compete in one of the
following top level market classifications:-
All those above are different
market segments.
And you can furthermore divide each of the above
categories into 2 parts -
open vs proprietary
architecture - which have different comfort zone acceptability with different
customers. And then again - another division will be by interface (FC SAN,
iSCSI, etc).
And
that's just one part of the enterprise SSD market which users have to think
about. I haven't even mentioned, above, subjects like
Vendors don't need to know all the answers - because
they can still sell SSDs easily anyway in this
growth phase of the
enterprise SSD market.
But users do have to worry about all
these problems - because if
they get it wrong and invest in dead-end architecture - then a few years
down the road they'll have to rethink their solid state storage infrastructure
all over again.
A facile answer to the question - what do enterprise
SSD users want? - might be to say - give users some degree of certainty
in a wildly uncertain world.
But that's not a practical formula either.
Is
the solution?
- better vendor communication about how they fit into the customer-centric
and application-specific SSD ecosystem?
(Efforts today in this
aspect are mostly laughable.) It could be all those things - but it
could be more too. Because maybe those aren't even the most important
questions.
As you can imagine - I get a lot of questions about the SSD
market.
And these help me to understand what I should be writing
about.
And as my readers include people in the SSD industry who make
new products happen - that feedback process helps moves things in some
directions which might be better than others. But it's not enough. And we can
all do better.
So here's what I suggest.
During the next few
quarters StorageSearch.com will be opening up the conversation re - what do
enterprise SSD users want?
My aim will be to collect a coherent set of
questions to design surveys which will inform, prompt and guide the SSD
industry about what it can do better - as seen from your point of view.
Is
this important? Does it really matter? Let me know what you think. | | |
| .. |
| the Problem with
Write IOPS |
the "play it again Sam"
syndrome
Flash SSD "random write IOPS" are now similar
to "read IOPS" in many of the
fastest SSDs.
So
why are they such a poor predictor of application performance?
And
why are some users still buying
RAM SSDs which cost
more than SLC and significantly more than MLC? - even when the IOPS specs look
similar. |
 |
This article tells you
why the specs got faster - but the applications didn't. And why competing SSDs
with apparently identical benchmark results can perform completely
differently. ...read
the article | | | |
| |
.... |
| "The need for fast
data erase - in which vital parts of a flash SSD are destroyed in seconds - has
always been a requirement in military projects. That's because if a disk falls
into enemy hands the data protection offered by encryption isn't safe enough." |
| ......from:-
Fast Purge flash SSDs | | |
| .. |
| Who's who in SSD? |
| .. |
 |
| OCZ, founded in 2002, and
headquartered in San Jose, CA, is a leader in the design, manufacturing,
and distribution of high performance and reliable SSDs. | |
| .. |
| "SSD Wear Leveling
is a technique used inside flash SSDs to prolong the life of a flash memory
array. Countering the phenomenom called endurance - Wear Leveling processes in
the SSD controller keep track of how many erase cycles have been performed on
each flash block - and dynamically remap logical to physical blocks using
algorithms which spread out the wear over the whole population in the array..." |
| ...from:- -
inside flash SSD jargon | | |
| .. |
|
|
| .. |
| "It's like putting the
wrong tires on your car. Everything's OK on the daily commute. You see the
difference only when you hit snow or drive around a bend at 120 mph." |
| ...from:-
endurance issues
in flash SSD operating life | | |
| .. |
|
|
| . |
|
|
| . |
| "A good way to
think about SSDs is like vitamin supplements or medicine for computers... You
have to be careful about swallowing any new pills just because you found they
got a good write-up on the internet. And it's the same with SSDs too.." |
| ...from:-
SSDs and tonic medicine | | |
| . |
|
|
| . |
| Surviving SSD
sudden power loss |
Why should you care
what happens in an SSD when the power goes down?
This important design
feature - which barely rates a mention in most SSD datasheets and press releases
- has a strong impact on
SSD data integrity
and operational
reliability.
This article will help you understand why some
SSDs which (work perfectly well in one type of application) might fail in
others... even when the changes in the operational environment appear to be
negligible. |
|
| | |
| . |
|
|
| . |
| an introduction to
enterprise SSD silos |
In today's SSD market there are many SSD
solutions which appear to solve today's problems economically - but which are
dead-end solutions from an SSD architecture point of view.
If you
install such systems now - then you will create future problems when you expand
your SSD infrastructure because those types of products and their suppliers may
not be around for too long.
Having a clearer picture of the
application silos into which all enterprise SSDs will fit from a market and
architecture viewpoint will help you navigate decisions safely. |
|
| | |
| . |
|
| |