Why Some Users Need ASAPs
In
an article published in
2003 StorageSearch.com predicted that server SSDs had the potential to
become a $10 billion / year market. I had used SSDs with multi-user servers in
the 1980s - and I surprised myself (as well as many SSD vendors) when I
extrapolated various readers and technology trends and saw how 4 factors would
intersect to create the massive new market for SSD server acceleration whose
growing pains we're witnessing today. Those were:-
1 - the relentless
and ever growing need for application performance
2 - the declining
growth rate in processor frequencies since the end of the 1990s due to the
physical limits caused by signal skew on fat data busses as they came out of the
chip. (You could more CPU cores in these chips - but you couldn't get faster
signals out.)
3 - the declining cost of semiconductor memory
4
- the brick wall in hard disk latencies which hadn't altered a jot since the 1st
15K RPM hard drives started shipping in 2000. Most industry analysts
never
expected to see 20K RPM drives. Although shrinking magnetic geometries
increased capacity and throughput - the random access times remained unchanged.
The
point at which users would turn to SSDs to speed up their servers would be
different in different markets - the user value proposition being the point
where it's cheaper to add SSDs than add new servers, or impossible to get the
same performance in any other way. These factors had been well known by a small
number of SSD experts for many years - but while memory prices put SSDs out of
reach - and while server oems could still sell fatter CPUs (with more cores)
rather than faster CPUs (with better peak performance) these solutions remained
the tools of last resort used by power users in the defense, intelligence,
broadcast and financial markets. And another factor stopping the secret getting
out was that customers who had got success in using these tools didn't want
their competitors (or enemies) knowing how they'd done it.
Now of
course the SSD server acceleration paradigm is common knowledge - helped to
some degree - by marketing hype from the
SSD notebook
market. And it's reasonable for users to ask the question - "If SSDs can
make our notebooks faster - why not our servers too?"
The answer
is - "Yes they can - but it may still cost more than you can afford."
If
your corporate data sits in 10TB, 100TB or multiple petabytes it's not
economically feasible to place all your data in an SSD. That's one difference
to the situation with notebooks. It's not just the cost of the memory - the cost
of the SSD controllers
rise astronomically as the speed goes up the levels needed to support servers
apps too.
Tactically - what server owners have done since the dawn of
the SSD server market is try to get as much acceleration as they can afford by
using as little SSD capacity as possible.
Traditionally in
RAID systems and
SANs that process
has meant analyzing the bottlenecks in server apps - where the most
IOPS occur
in the smallest identifiable part of the disk storage system. With some apps -
like databases - whole books have been written on the subject. It's still a
difficult and time consuming task - but if you can get hold of a good SSD hot
shot and the right analysis tools - the data hot spots can be migrated to the
fast SSD storage - and if you're lucky you can get 2x to 40x application
speedups by logically replacing a small percentage of your disk bound data.
And
that's where another economic factor comes in.
The supply of SSD hot
shots - engineers who really understand SSD tuning issues - is limited. Most
work for SSD vendors - and their time is best utilized by focusing on the
biggest new customer prospects. If you work in a user organization where your
SSD budget is less than a million dollars - it's unlikely you will get to meet
one of these people (unless you are willing to be a beta site for a new product
from a new company.)
So - for most of you - the options are:-
- become your own SSD tuning expert (risky unless you work in an environment
which encourages research and trial and error), or
- get someone who's an inexperienced SSD expert to do the job for you. (Ask
them how many SSD tune ups they did in 2005?), or
- test a product which does some of the tuning process for you. This is
the market niche for SSD ASAPs - which in my view is a very big market -
because most small and medium sized organizations don't have million dollar SSD
server budgets yet (December 2009) - even if their organization is already
deploying millions of dollars worth of servers.
There are many flavors
of "auto tuning" within the ASAP market space - and there will be more
to come. Currently they are segmented by interface type. In the future there
will be ASAPs which have been optimized for particular classes of application.
An important argument in favor of the SSDs ASAP type solution - is
that unlike a human tuned system - the ASAP should maintain its initial
effectiveness for longer - because it's always learning. Whereas in a
traditionally tuned system it may be necessary from time to time to revisit the
initial design assumptions if factors in the application environment change
considerably.
Important Warning! No matter how fast the SSD in
the ASAP - you will only get an economic speed up if the assumptions about data
use (designed into the box) correlate well with the actual frequency and shape
of the data usage patterns in your application. If that's not the case - then
shuffling data into a fast cache at a time when that particular data is not the
bottleneck - is simply an expensive way to achieve nothing at all. This is
something which you only learn by trial and error, and experience in modeling.
And another lesson I learned for myself in
1990 (which
is still true today) is that for some applications and some data sets - the
bottleneck is not the disk system but something else! And adding an SSD in
these circumstances achieves no speedup at all - even if all the data is sitting
on the SSD.
To summarize this important point - adding an SSD does
not make the application faster - if the data in the SSD is the wrong data (at
that time) or if the old hard disk system wasn't the bottleneck in the first
place.
I'm an evangelist for using SSDs in the right places for the right
reasons when they are economic. They have many advantages. But like any tool -
you have to know when its use is appropriate.
I just wanted to get that
out of the way. It's an important sanity check. Now let's assume that you might
have an application which might speedup economically using SSDs. You're trying
to investigate this subject in more detail and will do testing on whatever you
shortlist on a try before you buy basis (which the SSD industry started to
adopt more widely after seeing results from our
2004 SSD buyer survey).
Deciding
if you are an ideal user who should be looking at the SSD ASAP market (or
not) sounds like a complicated process. But I think there are some simple
filtering questions you can ask yourself - shown in the table below - which
might be helpful. |
What type of SSD
server acceleration tuning should I be looking at? - © 2009
StorageSearch.com
Which of the these best describes your application? |
thousands of servers
homogene ous
environment
Google style architecture. All servers have about the same
weight and run the same or very similar apps. |
Congratulations! Your budget is big enough to
attract an SSD hot shot.
They will probably steer you to embedded SSDs
(either PCIe SSDs or
2.5" SSDs
integrated in each server rack. |
thousands of servers
heterogeneous
environment
Command and control style architecture. Some servers are
heavy weight, others are light weight. They interact at many points in many
complex ways. |
Congratulations! Your budget is big enough to
attract an SSD hot shot.
If your low end servers are bottlenecks -
those may benefit from embedded SSDs (like above).
But that also
places more IOPS
stress upstream - where the only solutions may be
SAN (or
Infiniband)
compatible RAM SSDs.
In addition to the embedded SSDs downstream. |
tens to hundreds of servers
any
environment
any architecture. |
This is a gray area where you may or may not
attract an SSD hot shot.
At the top end of this range you will most
likely benefit from some kind of human adjusted tuning - because ASAPs are an
expensive option when scaled up (compared to the alternatives.)
At
the bottom end of this range using ASAPs (if they work for your type of
application) may be worth testing - because the cost of analyzing and tuning
your system using a human SSD expert may outweigh the theoretical cost
differences between an ASAP and a dumb vanilla SSD. |
1 to 10 servers
any environment
any
architecture. |
It's almost certain you won't attract an SSD
hot shot. (Unless you want to be a beta site.)
The good news is - there
are many possible different ways in which you could use SSDs (inside the box,
outside the box etc). The bad news is there are so many possible solutions which
might work too.
The entry level price for some types of ASAP may be the
determining factor which rules them in our out. Performance scalability is not
an important issue for this low server count. Total disk capacity may be.
You
may have to become your own SSD hot shot expert. | |
In the current state of the
market - with only a handful of vendors offering genuine SSDs ASAP products -
there is a limited range of choices for users who have any particular
interface preference. But I expect this to change in 2010 as the scale of the
market opportunity becomes better understood. This will be driven by the
growing gap between trained SSD hot shots and demand for SSD acceleration.
I'll
be adding more articles on this subject in the new year. |
 | |
| SSD ASAPs (and not quite
ASAPs / "pretenders") news extracts from recent
SSD Market
History |
|
FalconStor tunes
Violin's SSD
Editor:- March 2, 2010 - FalconStor today
announced
technical
and VAR channel support for Violin Memory's 2U
rackmount FC flash SSD
- the Violin 1010 .
Although
the headline specs of this very fast flash SSD are substantially the same as
when it was launched in
November 2008
the 2 important things which have changed are:-
- the price point
- $32,000 for the 500GB (lite capacity) version, and
- the availability of SSD
ASAP-like features implemented by FalconStor's SafeCache and HotZone
software.
Tiering SAN Shifts Real Estate without Costly Tears
Editor:-
February 22, 2010 - Compellent
published a
case
study (pdf) - which shows the benefits of
automated
tiering
SAN storage - applied to
the online marketing of real estate.
Demonstrating the flexibility of
Compellent's "Fluid Architecture" their customer -
WhereToLive.com - is quoted as saying
- "With the Compellent system... I'm able to get a million-dollar SAN
over time and without that one-time million-dollar capital expenditure."
What
is Fluid Architecture? - Compellent's VP of marketing, Bruce Kornfeld,
explains...
"Compellent'sFluid Data storage
enables automated tiering at a granular level between any drive technology,
speed and even RAID level.
Shifting data between SSD,
FC,
SATA, and
SAS works quietly
and unobtrusively in the background. Businesses want a "set it and forget
it approach" and that's why automated tiering has proven popular –
because it saves customers a lot on disk drives, space and power costs. The fact
that most large, legacy storage vendors are now introducing their own solutions
only validate that customers are asking for automated tiered storage. Automatic
tiering is one party no storage vendor can afford to miss."
Editor's
comments:- this month is the
8th anniversary of
the "Affordable SAN Initiative." Like
$$Ds - there's
affordable and AFFORDABLE.
Solaris, SSDs and Sun-Oracle - past failures - future
challenges
Editor:- February 3, 2010 - in a new article today I
look ahead to the
next 5 years of Oracle,
Solaris and SSDs.
I also look back and give you my
list of Sun's biggest market successes and failures. ...read the article
Avere Adds SLC SSD Options to 2U ASAPs
Editor:-
January 26, 2010 - Avere
Systems today
announced
it's shipping new
SLC
flash SSD options in its
FXT Series
10GbE NAS compatible
SSD ASAPs.
The
2U Avere FXT 2700 appliance (from $82,500) features 64GB of DRAM, 1GB of
NVRAM, and 512GB of SLC flash SSD. FXT clusters can scale to 25 appliances and
support millions of operations/sec and tens of GB/sec throughput.
"One of the main assumptions of Demand-Driven Storage is that
data access requirements are different across applications," said Ron
Bianchini, President and CEO of Avere Systems. "Applications that produce
heavy random read workloads are best addressed by SSDs and the FXT 2700 is
Avere's answer for those users who have a high-end NAS infrastructure that under
delivers when it comes to these types of applications." |
|
| EMC Casts
SSD Divining Rod into Hard Disk Arrays |
Editor:- December 8, 2009 -
EMC today published
a report on its new
fully
automated storage tiering concept which the company says will simplify user
operations needed to optimize storage allocation between
hard drives and
SSDs within the company's
arrays.
The company says some of this functionality is
now
available on some models.
Editor's comments:- although
better than nothing - adding a software manager retrospectively to storage
arrays which were never designed for SSDs in the 1st place can never deliver as
much performance as a true native
ASAP SSD appliance
(where some of the support is built into the hardware) - and nowhere near as
much performance as the
fastest SSDs (EMC
has never been in this list) when optimized by human SSD hot shots.
In
order to get the full benefits of the SSD acceleration paradigm EMC will need to
dump its legacy storage array designs and start offering boxes which have been
designed from the outset to support large amounts of
PCIe SSD capacity.
Without that - its systems will remain moderate performers at immoderate
prices.
To put it another way - bolting SSD tiering onto controllers
designed for hard drives is like trying to do air traffic control by having a
traffic cop standing on the ground and waving his stick. You can make the stick
a brighter color and give the pilot stronger glasses - but it's not going to
give you the traffic movements you get from integrated avionics.
Symantec Adds SSDs to Storage Migration Classes
Editor:-
December 7, 2009 - Symantec
announced an upgrade to its Storage Foundation management software which
enables it to
automatically
discover SSDs from leading vendors and optimize data placement on SSD
devices transparently.
Editor's comments:- this is a tool within the context of a
complex and expensive
data
migration service - rather than an auto-tuning SSD acceleration tool.
You'll still need the SSD Hot Spot Engineer to tell you where to migrate those
files to. As we head into
2010 - Year of the SSD
Market Bubble - you're going to see the word "SSD" appearing in a lot
of press releases from
software vendors as a
way of making their products sound sexier.
Panasas Supports SSD Layer in ActiveStor
Editor:-
October 21, 2009 - Panasas
has announced support for
SSD acceleration
within its Series 9
ActiveStor
hybrid storage systems (ASAPs).
A single 42U rack configured with the new Series 9 system is capable
of delivering an estimated 80,000 NFS operations per second, as well as 6
gigabytes per second of throughput.
Avere Launches Hybrid NAS SSD Rackmounts
Editor:-
October 5, 2009 - Avere
Systems unveiled its
FXT Series of
clusterable 2U rackmount
hybrid
NAS appliances.
Each
module contains upto 8x 3.5"
SAS
hard drives, 64GB
DRAM and 1GB of
nv RAM. The embedded
Avere OS
provides storage acceleration by dynamically tiering between the internal
rotating and solid state storage. List pricing starts at $52,500.
"The FXT Series is a milestone in the evolution of storage
products with its dynamic use of storage media to maximize speed while
minimizing cost," said Ron Bianchini, co-founder and CEO of Avere Systems. "The
end-result is a product line that can deliver tremendous business value to
customers by providing high performance and high efficiency to the storage
network simultaneously."
Editor's comments:- Avere is the
3rd company in recent weeks to announce an automatic solution for the age old
problem of accelerating
legacy hard disk array applications with solid state storage. There are
some interesting differences in approach and target markets.
Avere's
product is aimed at NAS
systems. It's a complete end user solution which includes the hard disks
which are to be accelerated. Avere says the new product can be configured with
upto 1.6TB of DRAM per cluster.
Dataram's product is
aimed at SAN systems.
It's an end user upgrade solution which fits between the customer's
FC switch and
pre-existing SAN rotating storage arrays. In some cases where users have already
over provisioned hard disks - the
XcelaSAN
may also, as a side effect, increase the usable storage capacity as well as
speed up the apps.
Adaptec's
product is aimed at DAS
systems. The
MaxIQ
SSD Cache Performance Kit is an integrator / oem solution which
simplifies the task of building a hybrid storage pool.
Key questions
for customers are going to be:- Does it work? How does the price / performance
compare to vanilla SSDs and human tuning? And how
reliable are the
new products going to be? Understanding the
failure modes in
large SSD arrays is not something that traditional storage designers know
very much about.
Dataram eliminates waits for the SSD Hot Shot / Hot Spot Engineer
Editor:-
September 28, 2009 - Dataram
launched the
XcelaSAN
- a fast 2U
rackmount flash SSD with 450,000 random IOPS performance (assuming 50/50
R/W and 4k blocks), and upto 8x 4Gbps FC ports - aimed at the
SAN application
acceleration market. Pricing starts at $65,000 for a unit with approx 360GB
internal flash, of which 128GB is effectively used as a cache.
"It
is now well understood that the benefit of a solid state infrastructure for
compute-intensive environments is higher application performance with less
equipment and lower operational costs," said Jason Caulkins, Dataram Chief
Technologist. "The question is no longer 'How can I benefit from solid
state storage?' but 'How do I best implement solid state in my existing
infrastructure?' With XcelaSAN, we enable organizations with performance
intensive applications to seamlessly add a dynamic, intelligent solid state
storage tier to their existing SAN environment."
Editor's
comments:- At 1st glance this product looks like many others which have
aimed at the traditional market of SAN users. But its revolutionary design opens
a new market which has been inaccessible to traditional
FC SSD vendors.
Dataram's product includes proprietary software - which does away with the need
for an SSD expert engineer to identify hotspots and relocate critical data. The
company says the XcelaSAN will automatically learn and self optimize during the
1st few hours of operation - and it will maintain application speedups even
when applications and loads change - which is not possible with human tuned
systems.
The search for a self
tuning agnostic
SSD software layer which sits between a SAN server and conventional rotating
disk bulk storage has been the Holy Grail of SSD oems for over a decade. None
have actually achieved it - till now. Although many vendors have developed
semi-automated tuning kits and strategies for common applications - they require
considerable expertise on the part of the applications engineer to make them
work well. That has slowed down the adoption rate of SSDs in many midsized
organizations which don't have a big enough installed base to attract the start
SSD talent to look at their problems. And it's also why SSD accelerators, have
not been viable as a reseller product.
When I spoke to Dataram's CTO,
Jason Caulkins, I was impressed by the depth of marketing thinking behind the
new product launch.
Dataram realized that simply launching a me-too SSD box would have an
uncertain outcome in a market that's already so crowded. And Dataram's corporate
memory goes back over 30 years to pioneering SSDs for minicomputers which
they launched in
1976. But
all memory companies know that in the future SSDs will use more memory than
traditional markets - such as server or pc motherboards. So it's important to
stake out ground in the SSD market.
I asked - where did the technology
come from? Jason said some of it came from Dataram's acquisition of
Cenatek - where he had
already been thinking about the SSD business model problem for many years. With
much bigger resources available after Dataram's acquisition - he's had teams of
software engineers working on the XcelaSAN concepts and licensed essential glue
where needed.
Will it work? Dataram says the XcelaSAN has been tested
and working in customer sites. Product shipments in the US start in the next
quarter. And the product is storage agnostic - meaning the customer can replace
their SAN arrays at a future date and retain the acceleration speedup. XcelaSAN
seems to offer a viable route for mid-budget user enterprises - who have
been neglected by SSD vendors for economic reasons - to join the march of the
SSD Revolution.
Is it competitive? - If you use my quick and dirty
magic number for SSD sever accelerators - (write IOPS divided by cost per TB) -
it's in the same order of magnitude as leading PCIe SLC flash SSD cards - so
it's definitely worth a look.
Adaptec Enters the SSD Market
Editor:- September 9,
2009 - Adaptec
announced a new platform for integrators building
hybrid storage
pools using SSDs (ASAPs).
Its
MaxIQ
SSD Cache Performance Kit (which operates with upto 4x customized 32GB
Intel SSDs) includes
software that identifies frequently (hot) read data blocks and optimizes
subsequent "reads" by moving "hot" data directly into the
SSD cache for lower latencies and higher system performance.
Adaptec
president and CEO Sundi Sundaresh said that the new product "Underscores
the potential that we see for significant future management and conditioning of
data through the I/O path, which is central to our new ... strategy."
Nimbus Makes "SSD Acceleration ASAP" a Breeze
San Francisco, CA -
April 2, 2008 - Nimbus Data Systems today unveiled its Breeze Hybrid
series of multi-protocol 10GbE IP storage systems.
Nimbus claims
this offers 2.5x the throughput and 75% lower cost than a 4Gbps
Fibre Channel SAN.
The Breeze H-series runs Nimbus' HALO storage operating system which combines
storage virtualization and advanced data protection and supports an
SSD acceleration option.
...Nimbus profile
Editor's
comments:- Nimbus carries on the torch of a network storage operating system
- which under the name "Cloudbreak" - was first developed by Nimbus's
founder at TrueSAN
Networks
.
That kind of groundwork thinking may help to make an SSD
work economically as part of a hybrid HDD-SSD accelerated array - while avoiding
the traditional high costs of manual setup and tuning. Like space-time - cost
is relativistic and depends if you're in the right dimension. A
fully-configured Breeze MH860 - with all software, 34TB of storage, and 64GB of
mirrored SSD starts under $120,000. | |
