Storage
Market Outlook 2010 to 2015.........................................
by
Zsolt Kerekes,
editor - November 12, 2009 |
If you know what's
going to happen in the next 5 years - it's so much easier to prioritize your
plans for the year ahead.
This is a time of year when many readers are
thinking about their storage
marketing plans for
2010.
This
planning process takes place against a background of long range assumptions
which are more confusing than at any time since
September 11, 2001.
I'm often asked for my views by analysts and vendors. Here are some
ideas for you to think about.
- Even if there is a recovery in the US market in
2010 - many
segments of storage will decline in the long term and disappear entirely. Yet
despite the extinction of many old storage species - overall new storage will
expand to become the biggest part of overall hardware budgets in IT, and will
replace processors and OS's as the most critical determinator of personality
in many consumer electronic products too.
- 50% of the (enterprise)
hard disk market will no longer exist in 5 years, and
none
of it will exist in 10 years (except in museums). But in the short term some
segments (related to
disk backup and high
capacity video content archiving) will grow in revenue.
In
2010 some
HDD makers will report high revenue growth (compared to the year before
period) - but if you compare these to revenues 2 years earlier - these will
be seen to be shifts in market share - and it's likely that the total HDD
market reached a historic revenue peak in 2008.
- Fibre-channel storage
will remain viable for another 5 years propped up by 16GFC - but this market
will continue declining in market share compared to other types of
NAS - and FC SAN revenue
for rotating storage arrays has probably plateaued already. Despite that - some
segments in the SAN market will continue revenue growth - in particular the
SAN SSD market - mainly
as a service to
maintain the usable life of already installed legacy HDD based SANs.
- The high end of the
RAID controller market
is going to disappear.
There's little point in spending money aggregating
IOPS in an
array of hard disks -
if the result costs more, is slower and is less
reliable to
operate.
Some oems, like
EasyCo, have been
marketing arrays of COTS 2.5" SSDs which use traditional RAID
controllers since 2007.
But this "open" type of design approach has been a
niche within the
SSD rackmount market. Controller companies need high volume markets. Within the
server accelerator part of the SSD market there are too many performance and
design compromises involved in keeping the SSD and RAID controller as separate
sub-systems. So fast RAID controller companies have to integrate SSD
functionality - or lose this market.
AMCC 3ware (later acquired
by LSI) was the 1st
RAID controller to publicly confirm it was working on SSD related products - in
May 2008. This
was in response to StorageSearch.com asking all the leading RAID companies what
their SSD plans were. Most of them - at that time - didn't seem to realize that
the SSD market would soon intersect with their businesses and render their
fast controllers irrelevant.
Viking was one of the 1st
oems to include a true SSD (with wear-leveling) as a module for RAID adapters in
early 2009. But
such solutions merely improve a product type within a declining market -
rather than extend the market's life.
SSD ASAPs - Auto-tuning
SSD Accelerated Pools of storage - represent one way for traditional RAID
companies to leverage their technology assets garnered from their hard drive
experience - while making inroads into the SSD space. It's a new market in which
there are no established market leaders.
PCIe SSDs are the
quickest way for HBA companies to enter the SSD market -because of the low
technical entry costs. And in January 2010
LSI
announced
its intentions to enter this market segment - making it approximately the
163rd company to enter the
SSD market.
- The optical storage
market will continue to spawn exciting announcements about comebacks as
a mainstream storage media - based on what new diodes can do at different
combinations of wavelength). But in a 5 year timescale optical drives will cease
to be offered as standard peripherals on PCs and will decline in relevance in
consumer markets too.
- SSDs, based on at
least
5
distinct memory technologies aimed at different application segments, will
continue to surprise, confuse and exasperate users and analysts - by having
members in the SSD product population which at the same time include the most reliable and
least reliable
products in the computer industry. The overall revenue for SSDs in 5 to 10 years
will exceed the historic peaks for all type of magnetic storage - as SSD
acceleration products
cannibalize server CPU,
software service and utility budgets.
|
|
|
|
Later:- June 4, 2010 - a reader said he
couldn't see how SSDs would get below high capacity HDDs in TCO - although I
thought I had explained the exact technical problems which needed to be solve
clearly in my article -
this way to the
Petabyte SSD.
This was my reply.
If you look at what
has been achieved
by the 2 storage technologies in the approximate decade 2000 to 2009 (at which
point SLC flash SSD could deliver the same capacity - 2TB - in a hard disk
form factor)
Hard disk capacity grew 5x
SLC SSD capacity- grew 71x
The magnetic storage media has been on a shallower capacity
increase curve than the semiconductor curve. And the semiconductor one can
continue at the same fast rate.
The problem when you get to Petabyte storage densities in a
relatively small form factor is that magnetic technology cannot deliver useful
products.
Magnetic technology would either melt - or only be feasible with
access times (like tape libraries) which are 30 to 60 seconds. (Ignoring the
more fundamental problem that magnetic recording technology can't scale to these
small geometries.)
In contrast SLC flash technology only needs evolutionary
changes to support Petabyte disks - and can do so at access times which are
useful in storage systems.
At the Petabyte disk level the cost of the electrical and cooling
power of HDD technology means that - even if the HDDs were free - the running
costs would make them higher than solid state.
If you doubt that Petabyte disks will be in the storage roadmap -
just reflect on how storage needs have grown in the past 50 years. And many
existing technology trends and technology trends will accelerate the creation
of user data.
| |
Past Long Range Storage
Predictions
Here are some earlier predictions I made about
the storage market.
1998 - I predicted that the storage market
would become an important and distinct market in its own right. That's when I
launched StorageSearch.com - which was the 1st
storage publication to
cover the market as a whole. Hundreds of others followed in the next few
years.
2001 - in an article called -
the Next Decade in Storage
- I foretold the reasons why
disk backup would
eventually replace tape backup.
A market process chronicled
here.
2003
- in an article SSDs a $10
Billion Market? - I explained why I thought SSDs had the potential to grow
into a huge market - when users recognized their potential to accelerate server
applications.
2004 - in an article called -
Predicting the Long Term Future of Hard
Disks, Tape and Optical Storage - I said "removable optical storage -
the descendants of today's DVD drives - may become obsolete."
2005
- I published the
SSD Market
Adoption Model - which predicted all the major markets in which SSDs would
be adopted from notebooks to servers in future years - due to user value
propositions which had nothing to do with cost per GB of storage capacity.
Most
other analysts and publications at the time didn't understand what I was talking
about. But many SSD companies did understand - and these articles on
StorageSearch.com had a major effect on how the market developed in the years
which followed their publication.
Also in 2005 - in an article
called 7 Years of
storagesearch.com - looking back / looking ahead - I predicted that
storage reliability
would become a major factor in the market by 2012. We've already seen in 2009
confidence eroding in the reputation of some manufacturers due their track
records of shipping flaky
SSDs.
In 2010 we're going to see the 1st Fizzings in the SSD
Bubble - as some oems rush to market anything they can ship to avoid
the risk of losing out - in what could become a $20 billon / year market within
the next few years. We will see a lot more stories about products which didn't
work as expected - and may even see whole businesses collapse or suffer badly
due to the mis-application of some kinds of SSD technology. "Reliability"
- which seemed like such a boringly safe assumption in the past - will become a
key factor in future spending plans in the 2011-2012 time-frame. | |
| . |
 |
| . |
|
|
. |
 |
. |
"With new software and architecture - a single petabyte of enterprise
SSD could replace 10 to 50 petabytes of raw HDD storage in the enterprise -
and still get all the apps running faster." |
|
| | |
. |
 |
. |
| I only write about SSD
related topics that I'm thinking about or am interested in. So the last thing I
ever thought I'd be doing in the last quarter of 2012 is writing yet another
article about hard drives - especially in this slot - the home page of
StorageSearch.com - which has been reserved in recent years for thought leading
SSD articles. |
| How will the hard
drive market fare... in a solid state storage world? | | | |
