|
| ... |
| why does
TMS still favor SLC for enterprise accelerator SSDs? |
Editor:- in December 2010 - I interviewed
Jamon Bowen,
Director of Sales Engineering for Texas Memory Systems
and asked him about the use of SSDs in financial applications like banks and
derivatives traders.
TMS has been selling high speed enterprise SSDs
longer
than anyone else in the industry - and for years has been the highest
ranking rackmount SSD
company in StorageSearch.com's quarterly
Top 10 SSD Companies List.
I approached the company to ask about 2 issues which occupy a lot
of thinking in the top end of the SSD market:-
- what's the role of MLC flash versus SLC, and
- what's happening in the market with RAM SSDs?
I started by saying
that among enterprise SSD makers - TMS now seems like one of the few
companies which doesn't include MLC in their server SSD product line.
Leading
the MLC enterprise tide -
Fusion-io has been
pushing its MLC based PCIe
SSDs into server apps since
2007. And
and since 2009
SandForce has
been offering its SSD oem partners (now numbering over 20 companies) an
easy to
integrate controller
technology which manages
data integrity
and endurance
to a degree which enables
intensive write
IOPS SSDs using MLC - to last for 5 years. And more recently - in September
2010 - a year after shipping its first enterprise SLC only SAS SSDs -
Pliant Technology
reacted to market demand and started offering customers MLC variants.
So
I asked Jamon Bowen - what's going on? What's TMS policy about MLC SSDs?
Bowen
said I was right in thinking that the company doesn't ship any MLC SSDs. Its
market is at the high end of the SSD performance range. As a company - he said
TMS has evaluated and used more types of memory technology than anyone else in
the market. And they have also evaluated consumer grade MLC.
He
mentioned several related reasons for not using MLC right now in their SSD
products.
He said current consumer grade MLC nand flash has endurance
on the order of 3,000 write cycles. Although it would be possible for TMS to
design an ASIC controller with heavy use of over provisioning and parallelism
(which would cut down the average latency and extend SSD life) the cost of doing
this - at the high end of the performance spectrum (and in the
legacy SSD market
where the flash management is transparently offloaded from the host) - would
nibble away at the native cost advantages of the cheaper MLC. He said the end
result would be a rackmount MLC SSD that was only 50% cheaper than using SLC.
And another problem would be that the company's burn-in process (done for QA as
part of manufacturing) would use up 10% of the endurance life before the SSD
even reached the customer.
Bowen said that new types of so called "enterprise
MLC flash" - which use enhanced cell programming circuitry in the chip
design - might shift the balance of the viability equation. He said that
customers would not see consumer grade MLC based SSDs from TMS in the next
couple of years. On the other hand he said that TMS is interested in eMLC and
its engineering team can move quickly to bring products to the market.
I said I was surprised by
how popular -
articles about flash endurance - had become with the readers of
StorageSearch.com - considering that readers have to get engaged with
semiconductor and
architectural
complexities to follow the arguments.
Jamon Bowen said if users were
fixated too much on wear out - they might not realize that other factors
in SSD design could contribute just as much or more to SSD failure.
On
the subject of SSD
technology education - in a follow up email Jamon Bowen suggested - "It
might be a worthwhile endeavor to start educating the reader base on the
difference
between eMLC and MLC as I expect that across the market vendors will be
highlighting eMLC vs MLC in a similar way that SLC and MLC are compared today."
Later:-
In December 2011 - I spoke to
Holly Frost, CEO
and Dan Scheel,
President of Texas Memory Systems about their new
fault tolerant
SSDs, what's going on in the SSD market, and the philosophy that steers
the design of their SSDs. You can see details in the article -
"StorageSearch
talks SSD with Holly Frost." | | |
| ... |
 |
| ... |
|
|
| ... | |
... |
 |
Megabyte spent another
pleasant day
shooting the breeze about SSDs. | |
| ... |
 |
| ... |
| RAM vs
flash SSDs decision tipping point |
Editor:- in December 2010 - I interviewed Jamon Bowen,
Director of Sales Engineering for Texas Memory Systems
and asked him about the use of SSDs in financial applications like banks and
traders - a market which he said accounts for most of their RAM SSD sales.
The company which started in RAM SSDs
over 30 years
ago - now sells more flash SSDs than RAM SSDs (even though the product
brand for both types of SSD is
confusingly called
RamSan.) Bowen said that flash is 70% of their business.
Jamon
Bowen said that in many bank applications RAM SSDs are actually cheaper than
flash - because of the small size of the data. TMS still sell a lot of 16GB RAM
SSDs.
Production bank systems are typically shared by many hosts and
get a lot of write IOPS / capacity. To achieve the same
reliability and
latency with
flash would
require over
provisioning which would drive the cost up.
He suggested a simple
rule of thumb for intensive IOPS bank SSDs on the SAN
- < 128GB capacity - RAM SSDs cheaper
- 128GB to 4TB capacity - middle ground could be either - or determined by
other constraints
- > 4TB - flash SSDs cheaper
Jamon Bowen said that the analysis
side of operations in banks is different. That tends to have much larger data
sets and is more read than write intensive. In these apps - flash SSDs are
usually more economic. | | |
| ... |
|
|
| ... |
|
|
| . |
 | |
