 |
| ... |
3 storage media contenders which threatened flash SSDs in the
past decade
by
Zsolt Kerekes,
editor - May 19, 2011 |
Sometimes the future
course of a market is determined as much by what didn't happen as by what did.
The
SSD market's rock star
cult status today owes much to 3 important rival technology phantoms
whose appearance in the flesh was once eagerly anticipated by
admiring storage fans.
But true to their vaporish origins and maybe
even due to shyness they stayed behind the opera curtains. (Were they ever
really there? Some people claimed to have seen them.)
Let's start with
hard drives.
Seagate launched the
world's 15,000 RPM hard drive in
February
2000 - and for years after there was much speculation about who would
crank up disk rotation speed to the next level (and thereby speed up
random IOPS).
Holy holographic comets!
Another dazzling
prospect which distracted our attention at about the same time was the
allure of an entirely new market which never happened (despite eye watering
quantities of
VC money being pumped
into it) and that was the optical
storage industry - which - at regular intervals - newsflashed that it
going to replace hard drives - because its researchers could wiggle a bunch of
photons in a space smaller than you could squish a tiny magnet.
And
if that wasn't impressive enough they went on to say that the latest photonic
drive prototypes had more capacity than hard drives, would be cheaper to
make and would be 10x faster too. Who could resist tossing another
few hundred million dollars into the ghostwriters pot to fund the next
exciting episode of that story? (Sadly the script and the design details
were lost. They were backed up on a spare lab version of the photonic drive
which proved to be unrecoverable even by the best
forensic sleuths in
the business. So they cancelled plans for the book, the tv series and the
movie.)
And in case you're wondering why the photonic market phantoms
didn't back up onto the
cloud? - they didn't
think it was reliable.
Online backup and storage was a flaky market in those days with companies
disappearing faster than a magician's rabbit - who incidentally was my source
for the tail end of that optical drive story.
Back to the spin from
the magnetics.
Things had gone quiet.
In February 2007 I got
fed up with waiting for faster hard drives and published an article called -
Will
there ever be 20,000 RPM HDDs?
By then it was clear to many in the
SSD market that zero RPM drives (
alias SSDs)
would eventually terminate the market for so called "fast" enterprise
hard drives.
The growth of SSDs is a much more
complicated
story than the simple replacement of one type of storage by another. SSD
revenue is a hungry feeder - it prefers eating servers - which have more
protein - but when it's hungry it will nibble away at some hard drives for
roughage too. Eventually SSDs will
eat
all the hard drives in the datacenter and become a monster market in its own
right.
HDD makers aren't scared by SSD apparitions either
Big
hard drive makers
don't
lose too much sleep about those scary bedtime stories right now because
they think the SSD monster is still too small to digest them and besides -
unlike a leopard - an HDD company could
easily can change
its spots should such an SSD exigency occur.
You may be sure
that HDD makers have already done a lot of reading up on those spot
changing tricks (or are cooking up
some of their own)
because if you look at recent market projections from
SSD analysts and
add them together it's clear that flash SSD revenue will overtake HDD revenue by
2015. I think HDDs will limp on for another handful of years in consumer markets
and may even be given
away free.
When a computer industry supplier segment stops
making its product faster - and concentrates on making them cheaper - it
doesn't satisfy the same needs. That's why HDD makers have been cultivating
new markets in consumer gadgets in the past decade which aren't nearly so
demanding for raw performance.
In a few years time - when you look
back on the
history of
the SSD market you'll be surprised that the
hard drive market hung
on for so long.
But as you're reading about the
storage market today -
you know that navigating the SSD market safely to this bright solid state
future feels more like crawling through a mine-field.
A couple of
wrong steps can leave you worse
off than your were to start off with.
There's no armor which is
going to protect you on this journey where staying at home and doing nothing
puts you at serious risk of IOPS starvation.
Your best defense is to
read and understand more about this subject.
Knowing more is
better.
And if anyone tells you the SSD market is easy peasy - "click
here to add to your basket and all your problems are solved!" or "we
tested 10 SSDs to find which is the best for you!" - trust me they don't
know what they're talking about. Only you can decide what's best for you -
because no one else knows your needs and priorities.
Postscript
comments
The above text was published on the home page of
StorageSearch.com as my April 2011 blog.
But when it came to saving it
as a permanent article I realised that the subject deserved another strand of
discussion to make it historically accurate.
I've continued that flow
of thinking in the column on the right hand side of this page.
Below
are links etc to some other articles I've written on the subject of the SSD
market's future. You can see more content on this theme from me and others -
in the SSD market
analysts page. | |
| ..................................................................................... |
| 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 archive SSD.
It will
start to replace the last remaining strongholds of
hard drives in the
datacenter due to its unique combination of characteristics, huge storage
density, low running costs and operational advantages.
Bulk
storage SSDs will displace the last remaining hard drives in the enterprise
server market by 2020 - even if the price of a new hard disk drops to zero
and enterprise HDDs are given away free!
 The
new business and architectural models of the datacenter - how we get from here
to there - and the technical and problems which will need to be solved -
are just some of the ideas explored in this
visionary article. | | |
| . |
| the future of enterprise
data storage? |
Editor:-
the
future of data storage is the lofty sounding but aptly chosen title of a
new article published online in Broadcast Engineering
magazine.
It's written by Zsolt
Kerekes editor of StorageSearch.com
(that's me).
It's a completely new article which takes as its starting
point - storage market models and concepts from several futuristic
articles which have already appeared here on the mouse site - advances them
and integrates them into a single cohesive whole. |
 |
It will give you a clear
idea of how all the incremental changes you read about in
storage news pages will
add up to a different future - and the business reasons why. Sit back and
...read
the article | | | |
| . |
| 7 SSD types will
satisfy all future enterprise needs |
Editor:- March 27, 2012 - The enterprise SSD
market is complicated enough already but despite that - only 7 distinct types
of SSD classes are all that are needed to sustainably satisfy all the
architecture needs in the pure solid state storage data center of the future.
That's
the message of my new article -
an introduction to
enterprise SSD silos.
"SSDs aren't islands - their
data always comes from and goes on to other SSDs. Enterprise SSDs which have
been designed without any reference to how they will be segmented marketwise
and interact technically with other types of SSDs - in the mind's eye of the
customer architecture - will eventually fail in the market..."
If
you're an enterprise SSD user - you can use these classifications to judge where
all the current SSDs in your shortlist today fit within the scope of your
(maybe as yet unclear) future SSD architecture.
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.
If you're an SSD
vendor - then having a clearer picture of the future enterprise customer mix
will help you kill bad product ideas before they sink the business.
Better still - knowing where your good products fit in the model will help
you market them more efficiently... |
 |
and may help you recognize
which SSD vendors could be natural partners - because you don't compete for
the same SSD cave space. ...read the article | | | | |
... |
 |
"Don't be shy" said
Megabyte.
"Our readers would love to see you." | |
| .. |
| "You read a lot on
web pages like this about SSDs one day replacing hard drives. Hard drives have
been around for 56 years - but nothing is sacred when it comes to technology
wars and for various reasons flash SSDs might be lucky to reach their mid 20s." |
| ...from -
HDD news - Rambus aims
at replacement for flash SSDs | | |
| ... |
| nv memory alternatives to
flash? - (updated July 2011) |
| The 3rd "no-show" factor in
the 2000 to 2010 period was the so-called "flash SSD killer" - a
competing type of non volatile memory which might (one day) replace flash.
There have been many different technologies in this category - today called "storage
class memory". Here's a small selection of nv memory cry-wolf stories that
I picked out of the
storage history
archive. | |
| ... |
|
|
| ... |
| MRAM - the new Core Store? |
Editor:- June 10, 2003 - IBM and Infineon Technologies
today announced they have developed the most advanced Magnetic Random Access
Memory (MRAM) technology to date by integrating magnetic memory components into
a high-performance logic base.
Today's announcement could accelerate
the commercialization of MRAM, a breakthrough memory technology with the
potential to begin replacing some of today's memory technologies as early as
2005. MRAM could lead to 'instant on' computers, allowing users to turn
computers on and off as quickly as a light switch. |
 |
"MRAM has the potential to become the
universal memory technology of the future," said Dr T. C. Chen, VP Science
and Technology, IBM Research. "This breakthrough demonstrates that MRAM
technology is rapidly maturing and could fundamentally alter the entire memory
marketplace within the next few years."
By combining IBM technology with Infineon's expertise in creating very
high-density semiconductor memory, the companies believe MRAM products could be
commercially available as early as 2005. | | |
| ... |
| SAMSUNG Introduces New
Nonvolatile Memory - PRAM |
| Editor:- September 11, 2006 - Samsung announced that
it has completed the first working prototype of what is expected to be the main
memory device to replace high density NOR flash within the next decade a
Phase-change Random Access Memory. |
 |
The company unveiled the 512M-Megabit PRAM
device in Seoul today. More scalable than any other memory architecture
being researched, PRAM features the fast processing speed of
RAM for its operating
functions combined with the non-volatile features of
flash memory for storage,
giving it the nickname "Perfect RAM."
A key advantage in PRAM is its extremely fast performance. Because
PRAM can rewrite data without having to first erase data previously accumulated,
it is effectively 30 x faster than conventional flash memory. Incredibly
durable, PRAM is also expected to have at least 10x the life span of flash
memory.
PRAM will be a highly competitive choice over NOR flash, available
beginning sometime in 2008. Samsung designed the cell size of its PRAM to be
only half the size of NOR flash. Moreover, it requires 20% fewer process steps
to produce than those used in the manufacturing of NOR flash memory.
...Later:-
- it took Samsunganother
3 years to get the 512Mb PRAM into production. | | |
| ... |
| IBM Previews Racetrack
Solid State Storage
|
Editor:- April 10, 2008 - IBM researchers have
published
details about a new type of high density non volatile memory.
"Racetrack" memory is so named because the data "races"
around the wire "tracks". IBM suggests that in the next 10 years - the
new memory could compete with
flash at a much lower
price per gigabyte. The new technology uses magnetic domain storage but without
the high current needed by earlier solid state magnetic devices.
...Later:-
if you ever wondered how all those other non volatile RAM-like technologies -
which flash into the SSD news
pages briefly and then disappear for years - might fit into the
future SSD landscape
- an IBM/SNIA white paper called -
Storage
Class Memory - the future of solid state storage (pdf) extrapolates memory
chip technology trends upto 2020. | | |
| . |
| Unity Semiconductor Unveils
Flash's Successor |
Editor:- May 19, 2009 - Unity Semiconductor
exited stealth mode and stated its aim to have the lowest manufacturing
cost per bit in the non volatile memory industry with a new breakthrough
technology called
CMOx.
The
company said it will ship 64Gb devices in volume in 2011. Unity Semiconductor
says it will develop and produce NAND flash successor technologies and
products that, in time, will extend into high ]performance embedded and
enterprise applications.
"It's a Technology for Terabits that
will challenge high volume rotating magnetic media" said Unity
Semiconductor Chairman, President & CEO Darrell Rinerson a former executive
at Micron Technology
and at AMD.
The
company, also announced today it has closed a Series C funding round for $22
million. This brings to nearly $75M the total funding to date in Unity
Semiconductor. | | |
| ... |
| world's first PCIe PCM SSD |
Editor:- June 14, 2011 - NVSL ( the Non-Volatile Systems Lab at
UCSD) recently
demonstrated
a prototype PCIe PCM (phase-change memory) SSD - with R/W speeds upto 1.1GB/s
and 327MB/s respectively and 8GB usable capacity.
A spokesperson for
the Moneta SSD design team - Professor
Steven Swanson said "...Moneta gives us a window into the future of
what computer storage systems are going to look like, and gives us the
opportunity now to rethink how we design computer systems in response."
Swanson says he hopes to build the 2nd generation of the Moneta
storage device in the next 6 to 9 months and says the technology could be ready
for market in just a few years as the underlying phase-change memory technology
improves.
Editor's comments:- in a white paper
Protoype
PCM Storage Array (pdf) the team outlines the design and architecture of
their PCM SSD prototype and also compares aspects of performance with entry
level PCIe flash SSDs from
Fusion-io. In a
recent article
I warned that you should not pay too much heed to comparative PCIe SSD
benchmarks - because from different arbitrary selected angles they can "prove"
different arbitrary performance rankings. I wouldn't be surprised if some
investors take fright that a PCM SSD scored higher than a Fusion-io SSD in some
of the published graphs. But for those who understand SSD architecture it
doesn't reveal anything new.
In my view this prototype clearly
demonstrates the strengths and weaknesses of PCM as an SSD technology.
PCM SSD strengths vs flash
The granularity of writes
in PCM is smaller and faster - which means that small R/W operations have higher
IOPS. If you have apps where that is important you can simply buy
SSDs with various
ratios of integrated RAM cache. That would give you small block IOPS
better than PCM - end of story. PCM has higher
endurance
than SLC - which means that the
SSD controller
overhead applied to endurance can be lighter than in most flash systems. Hence
potentially faster latency through to the media.
PCM SSD
weaknesses vs flash
The prototype PCIe SSD card provides capacity
which is similar to RAM SSD
density - but with a large block R/W throughput which is much lower than
flash arrays. This
implementation used 16MB PCM chips.
Flash allows higher capacity writes
to a single chip - and this gives better peak performance results than PCM when
exploited in parallel architecture arrays. You can't get those flash peak
performance numbers from a PCM array in the same board footrpint - because many
PCM chips have to be written to concurrently to achieve the same capacity R/W
as a single flash chip. That means with today's technologies - flash SSDs
have a higher proportion of ready to write memory chips in the same chip count
population as PCM SSDs.
For more about alternative SSD technologies -
see SSD's past phantoms. | | |
| . |
| "PCM is a bigger
threat to supercaps than it is to flash..." |
...Editor talking
to various strategists and investors in June in response to comments about
a news story
by IBM that its researchers had proven the viability of multi-bit storage.
In the past decade PCM had lost several previous geometry
generations of economic viability because flash capacity roadmaps had gotten
several boosts from MLC and then x3 - each time raising the bar for PCM.
translation:-
phase memory may replace power holdup
components
and DRAM caches
in some flash SSDs. But the idea that pcm will significantly replace flash
in SSDs in the next year or so is ridiculous.
IBM and other non
volatile memory makers have been over hyping their technologies and saying
that they will kill or replace flash soon -
for the past 8 years.
And I can remember similar claims from other nv RAM researchers going back to
the 1970s (before flash). | | | |
