|
| .. |
the new SSD uncertainty
principle
resetting the assumptions about flash endurance
by
Zsolt Kerekes,
editor - July 25, 2011
Recently I had a long discussion with
STEC about how
they manage MLC flash inside their enterprise and industrial SSDs and what's
different about what they do.
To be honest I wasn't expecting to hear
anything new. It's over
30 years
since I came across the concept of wear-out in non volatile semiconductor
memories in my previous career as an electronics engineer.
And it's
over 6 years since StorageSearch.com published the first white paper
devoted exclusively to the subject wear leveling (written by SiliconSystems -
now part of WD).
Nowadays
rarely a week goes by without some mention of subtle factors in dealing
with MLC speed and reliability from memory chip makers,
SSD controller companies
and even systems companies.
But STEC is a
significant SSD company
so I thought I'd be polite and listen.
My first impression
during STEC's presentation was - this is just a comforting sounding bunch of
new brand names for techniques which are already well established and which
other controller companies have already been doing for years.
But I
was at first surprised and then delighted to see I was wrong.
STEC has discovered that MLC flash endurance is a more elastic
concept than most experts believe.
By using back doors into the flash chips and studying this
problem for many years STEC has discovered that some ways of managing industry
standard MLC result in much better endurance, sustained performance and data
reliability than simply relying on wear leveling protocols. It's the how you
write and not just how much you write that makes the new difference.
I
find it hard staying quiet long when people are talking about SSDs - so
pretty soon I was asking lots of questions. The main one being - why have
you been keeping this stuff secret? Everyone should know about this....
Soon
everyone will.
STEC has lifted the NDA on the presentation they gave
me - from which the images and highlights below are taken. | |
|
notes adapted from an STEC presentation about true enterprise MLC
SSD technology
- A growing number of enterprise applications require highly reliable,
high-performance, long-lasting solid-state drives at a lower cost than today's
SLC flash-based drives.
- Deployment of standard consumer-grade or so-called enterprise MLC (eMLC) in
SSDs for enterprise applications do not meet requirementsfor performance,
endurance, and reliability required.
STEC's CellCare™ Technology
Fills the Gap Between eMLC and SLC Flash
STEC says that so-called "enterprise
grade" MLC has weaknesses when used in enterprise apps.
- can't guarantee 10x full disk writes per day for 5 years - which is
a rule of thumb requirement for typical enterprise SSD storage. (Some SSD
storage in cache apps need much more endurance than this figure - but this can
be accomodated using
overprovisioning,
RAM cache or
SLC.)
- long term availability - eMLC is currently a marketing / process
tweak / selection from memory suppliers which provides
endurance at
an intermediate between SLC and consumer MLC. These tweaks may not work or
deliver usable enterprise MLC in future product generations - as geometries move
from 3x to 1x nanometer.
- data
integrity in eMLC worsens with wear - which can result in dramatic read
performance degradation after several years use in customer systems
- eMLC costs
more than consumer grade MLC. A technology like CellCare can potentially
deliver lower cost enterprise SSD capacity - particulalry when it's packaged
in an ASIC rather than FPGA.
|
 |
 |
STEC has been collecting data about flash cell data integrity,
performance and endurance for several years.
This led the company to
conclude that conventional wear leveling schemes don't work as well as its own
technologies. |
 |
For more info about STEC's CellCare™
Technology - take a look at
STEC's
whitepaper (pdf)
To see what other leading enterprise SSD makers
have to say about the issue of SLC vs MLC vs eMLC take a look at these articles.
| |
| . |
 |
|
| |
|
related articles
SSD efficiencies
Sugaring MLC for the
enterprise
adaptive R/W &
DSP IP in flash SSDs
Data Integrity
Challenges in flash SSD Design
SSD Myths and
Legends - "write endurance" |
|
.. |
|
"...flash
endurance is a much more
elastic concept than I had thought" |
...editor talking to STEC after
reading
about the results of their multi-year
flash data integrity
testing programs. | | |
|
.. |
|
 |
|
.. |
| .. |
| How big was the
thinking in this SSD's design? |
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 | | | |
|
.. |
|
|
|
. |
|
| the 3 fastest flash
PCIe SSDs - list / lists |
Are you tied up in
knots trying to shortlist flash SSD accelerators ranked according to
published comparative benchmarks?
You know the sort of thing I mean -
where a magazine compares 10 SSDs or a blogger compares 2 SSDs against each
other. It would be nice to have a shortlist so that you don't have to waste too
much of your own valuable time testing unsuitable candidates wouldn't it?
StorageSearch's long running
fastest SSDs directory
typically indicates 1 main product in each form factor category but those
examples may not be compatible with your own ecosystem.
If so a
new article -
the 3 fastest PCIe
SSDs list (or is it really lists?) may help you cut that Gordian
knot. Hmm... you may be thinking that StorageSearch's editor never gives easy
answers to SSD questions if more complicated ones are available.
|
 |
But in this case you'd be
wrong. (I didn't say you'd like the answers, though.) ...read the article | | | |
|
. |
|
| SSD Myths and
Legends - "write endurance" |
StorageSearch started lobbying flash SSD
makers to standardize on a way of specifying SSD endurance in 2006.
Back
then vendors were reluctant to talk about this issue because they were worried
that user fears about sudden SSD death would backfire on the industry - and they
each had their own secret ways of managing flash endurance.
Nowadays
you can't stop SSD vendors talking about how clever they are at dealing with
endurance. In theory the problems are now well understood - but solving them
presents a challenge for each new flash chip generation. |
 |
This classic article- seen by millions of
readers - has been updated hundreds of times. ...read the article | | | |
|
. |
|
 |
|
. |
|
| 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. |
|
| | |
|
. |
|
|
|
. |
|
|
|
. |
|
| Bad block
management in flash SSDs |
This is an introduction to the thinking behind
one of the many vital functions inside a flash
SSD controller.
Native
media defect quality in new flash memory chips has grown steadily worse in the
past 10 years as geometries have shrunk. |
 |
This article enumerates
the scale of the problem and explains how intrinsically dodgy flash memory
is transformed into dependable flash SSDs which you can entrust with your data.
...read the
article | | | | |
