 |
|
|
by Zsolt Kerekes,
editor - September
29, 2010 |
You may think that the so
called "enterprise
SSD" market is already complicated enough - but there are many more
product implementation choices and technologies which will make the spectrum of
enterprise SSDs clamoring for your attention wider than it is now and harder to
focus on and latch onto quickly.
I'm not just talking about the
proliferation of memory types. There are 5 different memory types currently
being marketed in "enterprise SSDs" and more on the way. |
And I'm not talking about the
different architectures (open versus proprietary)
which already occur in vanilla
rackmount SSDs.
And
I'm not just talking about whether users are better off with SSDs which can
tune themselves (using one of the many different types of algorithms embedded in
SSD ASAPs) versus the
old fashioned way of getting
SSD tuning done
manually.
And I'm not just talking about application specific SSDs -
which already exist in some market segments - and will be a much bigger part
of the future SSD market.
What I'm proposing is a simplistic way
to view all these products (and those still to come) through a
technology-agnostic lens (mental trick) which quickly helps you decide whether
you should invest more time reading about new SSD products which fall under
the "enterprise SSD" umbrella.
In my view all enterprise
SSDs can be thought of as belonging in one of 2 categories
- Legacy - the SSDs are going into an architecture originally
purchased by the user with just
HDDs in mind.
- New Dynasty - the SSDs are going into a user base or factory fitted
box always intended to have SSDs from the initial purchase order.
Despite
the names "Legacy" and "New Dynasty" - both types of SSDs
will both be around for the foreseeable future.
PCIe SSD examples
In
the
PCIe SSD market you
get both types of SSD. Some products are legacy while others are new dynasty.
And knowing which is which can help you comprehend some puzzling market
nuances.
You may think that products with superficially similar
performance envelopes like the
RamSan-20 from
Texas Memory Systems
and the ioDrive from
Fusion-io compete
head to head.
But that is rarely the case for most commercial end
users.
The RamSan-20 (which has an onboard offload
SSD controller) is in
the Legacy camp. It's a product which a user might retrofit to an existing bunch
of servers running existing apps.
The ioDrive is undoubtedly in the New
Dynasty camp - because it will usually be a factory fitted option supplied by
the server oem. Although you can use ioDrive's in some older servers - these
products work better with newer servers (with faster CPUs) because that's the
market they were designed for - with the
host CPU doing
the memory gymnastics traditionally done by an SSD resident hardware
controller.
The Legacy / New Dynasty way of looking at things works at
the next level up in the application hierarchy too.
A New Dynasty SSD
accelerated server (whether it's got a factory fitted PCIe ioDrive inside or
some SandForce driven
SAS SSDs) simply looks
to the external world like a faster server. A year or so after you've installed
a couple of hundred New Dynasty accelerated servers into your environment - you
may hit some new IOPS bottlenecks in another hot spot further up the data food
chain. In most user sites that bottleneck will be architected on a classic
storage network (SAN or
NAS) and the solution to
the next level of bottleneck will actually be a Legacy SSD - maybe even a RAM
SSD.
The RAM SSD
market (the oldest part of the enterprise SSD market) may - in my analysis
-paradoxically see a resurgence and continue growing for many years.
Although
lower priced flash SSDs killed off the entry level RAM SSD market - only RAM
SSDs have the fast latency and symmetric performance which can cope with the
increased data demands which will be created by bigger populations of flash SSD
accelerated servers.
And meanwhile - all that propaganda spewing out
from the flash SSD
market in the past 5 years has educated users and made the market more receptive
to the idea of an SSD solution.
In the
old days
of the enterprise SSD market - pre 2005 - the server accelerator SSD was
parachuted into performance distressed customer sites as a previously
unlooked for expensive fix when all else had failed. |
 |
I'll talk more about this
new way of looking at enterprise SSDs - in future articles. Below you can see
links to a small selection of the many enterprise SSD directories and
articles here on this site. | |
|
| |
| ... |
|
| ... |
|
|
| STORAGEsearch is
published by
ACSL | |
| sugaring MLC for the
enterprise |
When flash SSDs started to be used as
enterprise server accelerators in 2004 - competing
RAM SSD makers said
flash wasn't reliable
enough.
RAM SSDs had been used for server speedups
since 1976
- and in 2004 they owned the enterprise market. (Before 2004 - flash SSDs
weren't fast enough and had mostly been used as rugged storage in the
military and
industrial
markets - and in space
constrained civilian products such as smartphones.)
By 2007 it was
clear that the endurance
of SLC flash was more than good enough to survive in high
IOPS server
caches. And in the ensuing years the debate about enterprise flash SSDs shifted
to MLC - because when systems integrators put early cheap consumer grade SSDs
into arrays - guess what happened? They burned out within a few months - exactly
as predicted.
Since 2009 new
controller
technologies and the combined market experience of enterprise MLC pioneers
like Fusion-io and
SandForce have
demonstrated that with the right management - MLC can survive in most (but
still not all) fast SSDs.
Now as we head into 1X nanometer flash
generations new technical challenges are arising and MLC SSD makers disagree
about which is the best way to implement enterprise MLC SSDs.
Which
type of so called "enterprise MLC" is best? Can you believe the
contradictory marketing claims? Can you even understand the arguments? (Probably
not.)
And that's why marketing is going to play a bigger part in the
next round of enterprise SSD wars as SSD companies wave their wands and reveal
more about the magic inside their SSD engines to audiences who don't really
understand half of what they're being told. |
|
| | | |
... |
 |
Megabyte's enterprise
accelerator
had elements of both traditional and
futuristic design with
built-in
redundancy and seamless failover. | |
| ... |
| New Dynasty -
Fusion-io,
FlashSoft |
| Legacy - Texas
Memory Systems, Dataram,
Virident Systems,
Violin Memory |
Why do I have auto-tiering
SSD ASAP companies
(FlashSoft and Dataram) in both of the categories above? Isn't an
auto tiering SSD /
ASAP always going into a legacy storage systems?
That's mostly true
- but on balance I'm listing FlashSoft in the new dynasty category because I
think that most of its installations are going to be pre-integrated by the apps
server supplier.
And when I talked to the company about its
technology roadmap (June 2011) it's clear they're already looking at what
their software will be able to do in
100%
solid storage environments - in which they'll be tiering between different
SSD speeds. | |
| ... |
"...You
like the idea - SSDs could make your apps go faster.
Problem is - you're
not in an industry where you can stuff
raw low latency and high
IOPS in one end of your business
sausage machine and expect to see
increased revenue and
dollars streaming out the other end..." |
| ...Editor:- in
the need for auto tiering
SSDs / SSD ASAPs | | |
| ... |
|
|
| ... |
| S/e/brand X/LC -
flash wars in enterprise SSDs |
When flash SSDs started to be used as
enterprise server accelerators in 2004 - competing
RAM SSD makers said
flash wasn't reliable
enough.
RAM SSDs had been used for server speedups
since 1976
- and in 2004 they owned the enterprise market. (Before 2004 - flash SSDs
weren't fast enough and had mostly been used as rugged storage in the
military and
industrial
markets - and in space
constrained civilian products such as smartphones.)
By 2007 it was
clear that the endurance
of SLC flash was more than good enough to survive in high
IOPS server
caches. And in the ensuing years the debate about enterprise flash SSDs shifted
to MLC - because when systems integrators put early cheap consumer grade SSDs
into arrays - guess what happened? They burned out within a few months - exactly
as predicted.
Since 2009 new
controller
technologies and the combined market experience of enterprise MLC pioneers
like Fusion-io and
SandForce have
demonstrated that with the right management - MLC can survive in most (but
still not all) fast SSDs.
Now as we head into 1X nanometer flash
generations new technical challenges are arising and MLC SSD makers disagree
about which is the best way to implement enterprise MLC SSDs.
Which
type of so called "enterprise MLC" is best? Can you believe the
contradictory marketing claims? Can you even understand the arguments? (Probably
not.)
And that's why marketing is going to play a bigger part in the
next round of enterprise SSD wars as SSD companies wave their wands and reveal
more about the magic inside their SSD engines to audiences who don't really
understand half of what they're being told. |
|
| | |
| ... |
| How big was
the thinking in this SSD's design? |
Editor:- July 5, 2011 -
Does size
really does matter in SSD design? - By that I mean how big was the mental
map? - not how many inches wide is the SSD.
The novel and the short
story both have their place in literature and the pages look exactly the same.
But you know from experience which works best in different situations and
why.
When it comes to SSDs - Big versus Small SSD architecture - is
something which was in the designer's mind. Even if they didn't think about it
that way at the time. |
 |
For designers, integrators,
end users and investors alike - understanding what follows from these simple
choices predicts a lot of important consequences. ...read the article | | | |
| ... |
| this way to the Petabyte
SSD |
In 2016 there will be
just 3 types of
SSD in the datacenter.
One
of them doesn't exist yet - the bulk storage SSD.
It will replace the
last remaining strongholds of
hard drives in the
datacenter due to its unique combination of characteristics, low running costs
and operational advantages. |
 |
... |
The new model of the
datacenter - how we get from here to there - and the technical problems which
will need to be solved - are just some of the ideas explored in this
visionary article. | | | |
| . |
| the Problem with
Write IOPS in flash SSDs |
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 users still buying
RAM SSDs which cost an
order of magnitude more than SLC? (let alone
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 | | | | |
