|
Founded in 1967, Dataram
is a recognized worldwide leader in the manufacture of high quality computer
memory and a pioneer in auto tiering SSD SAN storage appliances. |
.... |
 |
The company delivers value to
its customers through solutions that optimize data center and application
performance, while at the same time delivering significant cost savings without
introducing risk. Dataram products and solutions are deployed in 70 of the
Fortune 100 companies and governmental agencies including the Department of
Defense who use Dataram products for the most demanding mission critical
applications. |
| |
| see also:-
Dataram
- editor mentions on StorageSearch.com and
Dataram's SSD blog
|
- editor's comments:- November 2011 -
Dataram markets
rackmount SSD
appliances in a market category which I call
SSD ASAPs (SSDs As
Soon As Possible aka auto-tiering SSD accelerators).
In my SSD
business model and architecture classifications - Dataram is
Legacy and
Big Architecture.
If you click those preceding links you'll see some interesting competitor
groupings.
Users in the emerging SSD ASAPs market have frustrated
many vendor expectations by
treading
slowly, cautiously and hesitatingly towards these products - instead of
rushing to adopt them. Users wanted to see if claims made by vendors
(that their products would automatically work to speed up apps) would be proven
in practise (preferably somewhere else). In the last few years Dataram, like
other ASAP vendors found that instead of getting a rush of customers beating a
path to their door - the company had to instead invest considerable resources to
acquire customer test and reference sites.
Dataram publicaly
acknowledged that its original XcelaSAN lacked HA features which were needed
to make it more appealing to users in the fibre-channel
SAN market. The company
said it had fixed this feature gap with the
announcement
of its XcelaSAN Model 100 in March 2011.
Competitors to Dataram?
Dataram
has few technically equivalent competitors - but one of those in the FC SAN
ASAP category is recently emerged
GridIron Systems.
Customers
who want to accelerate their FC
SAN storage - using tier 1 type products - and who have small enough total
capacity to place it all in SSD - might also consider looking at
Huawei Symantec ,
Kaminario and
Violin Memory.
If
you're happy about putting cards inside your servers, or replacing the servers,
while keeping your existing SAN storage, another option is to look at
PCIe SSDs in
particular those with auto-tiering software support - such as
Fusion-io. The PCIe
route won't work if your SAN data is the bottleneck for a bunch of distributed
servers - but it might work OK if your installation only has a few colocated
servers - which can be mopped up into a single faster SSD accelerated unit. |
|
|
|
In October 2008 -
Dataram re-entered the
SSD market with the acquisition
of strategic assets from Cenatek
whose CEO has joined Dataram to lead the company's return to solid state
storage, an area they "pioneered 32 years ago.." ...Earlier:-
in 1976
-
Dataram sold an SSD
called BULK CORE
which attached to minicomputers from
ModComp and emulated hard
disks made by DEC and Data General. Each chassis held
8x 256k x 18
RAM modules and had a capacity of 2 megabytes.
This is a very
significant milestone for the
SSD market because it shows
the strategic value that memory makers place on SSDs.
In the past
companies like Intel have
resold 3rd party SSD cards, STEC
divested itself of its vanilla memory business and
Samsung would like to
own and control MLC patents now in the hands of
SanDisk.
Look
at it from the viewpoint of a memory maker.
Future server systems
will have orders of magnitude more memory in the attached SSDs than installed
as main RAM memory. Who
owns the brand of the SSD boxes will mean a dramatic difference to attainable
revenue. Being locked out of the SSD box - will mean that a memory maker can
only access smaller markets - or supply other SSD oems at commodity proces.
I expect to report many more such
acquisitions during
the next few years.
When discussing this story my wife said this is an
example of a marketing concept
called "forward
integration."
In August 2009 -
Dataram said it will
launch an SSD
accelerator
at SNW in October. The product is
currently being evaluated by key customers.
"As we prepare to
launch a data storage acceleration product, we have studied the current state of
solid state storage appliances very carefully to understand the strengths and
weaknesses of available solutions," said Jason Caulkins, Dataram's Chief
Technologist. "The features, benefits and hidden compromises the customer
must accept with today's generation of solid state storage appliances are not
always obvious. We hope to help our customers understand all their options and
provide them with a much better solution."
In September
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.
In
November 2009 - Dataram
is running a webinar -
Navigating
the Maze of Solid State Storage Solutions. Viewers will discover - "How
to better gauge your storage traffic to identify bottlenecks and areas where
solid state storage can provide a day 1 positive ROI."
In July 2010 - Dataram reported that
its
annual
revenue for the year ended April 30 grew 70% to $44 million
incurring a net loss of $1.6 million. Dataram's president and CEO - John H.
Freeman said the company is increasing resources into evolving its
XcelaSAN (ASAP) product line and
plans to launch HA versions later in the year.
In August 2010
-
StorageSearch.com published
SSD Bookmarks
- suggested by Jason
Caulkins, Chief Technologist
Dataram.
In October 2010 -
Dataram launched a
campaign called -
Let
Us Prove It to You - to persuade users to evaluate its
XcelaSAN (an FC
rackmount SSD ASAP)
- which the company says can dynamically and transparently improves I/O
performance up to 30x in the users own SAN environment with
block-level caching. Dataram is so confident in the results and ease of
installation that a 32GB iPad will be given to organizations that complete the
evaluation.
In April 2011 -
Dataram has doubled the
RAM cache available in its
XcelaSAN
(2U rackmount fibre-channel SAN
SSD accelerator) to
256GB (the system price is approx $75,000). XcelaSAN delivers up to 30x
transparent R/W acceleration to attached disk storage arrays with a
high-availability architecture (internal performance is
upto
450,000 IOPS). Unlike most solid state storage solutions, XcelaSAN
dynamically caches high I/O activity application data when it is needed, to
support multiple applications many times larger than the cache itself.
In
September 2011 - Dataram
announced
that
Dell
OEM Solutions will manufacture and support Dataram's
FC SAN compatible
auto-tiering / SSD ASAP
- the XcelaSAN from
November 2011.
In October 2011 -
Dataram published a 25
minutes video describing
its XcelaSAN auto-tiering SSD ASAP. The interesting parts for me are in the
last 5 minutes when we learn a bit more about its latency, IOPS and HA
architecture. In my view the company should have published more details about
these aspects of the product in a white paper a year or so ago. System
architects need internal guides and modeling metrics to help them understand
the strengths and weaknesses. |
|
| |
|
| "The ASAP uses its own
intelligence to figure out data hot spots - and tries to ensure - that as often
as possible - the data which you want from the slower, bigger store - is
accessible to your servers." |
| ...from the article -
the New Business
Case for SSD ASAPs | | |
| ... |
. |
|
 |
|
... |
|
| will rental
break through the indecision barrier for SSD ASAPs? |
Editor:- January 26, 2012 - One of the business
development obstacles facing
enterprise SSD ASAP
/ caching vendors in the past few years has been that users have mostly
thought of them as being HDD array accelerators.
And even if a user is
interested right now - and even if they are happy with their try before you buy
results - they often hold off making a purchase - because they think (after
reading web sites like this one) that one day they'll be ripping out their
rotating RAID systems and
replacing them with SSDs - so it might be silly to buy an SSD cache appliance
right now - if it only speeds up HDDs.
Now in reality - most users
won't replace their entire HDD storage as quickly as they might like to think -
and ASAPs do have a permanent role in the pure SSD datacenter too. Some
vendors' marketing materials talk about that - while others are still
harping on
about hard disks and the "superiority" of SSD - even when their
technology roadmap works just as well for SSD. |
|
| | |
|
. |
|
|
|
... |
|
| Dell
will distribute Dataram's auto tiering SSD |
Editor:- September 22, 2011 -
Dataram
announced that Dell
OEM Solutions will manufacture, provide hardware customization,
distribute and support Dataram's
FC SAN compatible
auto-tiering / SSD ASAP
- the XcelaSAN from
November 2011.
Editor's comments:- Since Dataram launched the
XcelaSAN 2 years ago
it has fixed perceived gaps in its failover characteristics and established
some impressive customer reference sites. But sales have been slow.
Part
of the problem has been that this product is aimed at users who don't have
the technical resources within their workgroups to
tune vanilla SSD
accelerators in SANs because of the
many complex data
architecture decisions which then arise. That's why they need
auto-tiering.
But without internal safety nets these ideal potential
customers have to be absolutely confident that it works and will be
supported. |
 |
This deal with Dell goes a
long way to doing that - and will tip the balance for many who liked the idea
but needed the reassurance that a 3rd party heavyweight company has looked at
the design and is prepared to support it. | | | |
|
... |
|
| the new SSD uncertainty
principle |
"enterprise MLC SSDs" - aren't all
created equal
Editor:- July 25, 2011 - STEC is starting to
lift the veil off how it manages MLC flash inside its enterprise and industrial
SSDs. You're thinking - isn't it all the same? Just a variation on what
SandForce and
Fusion-io already do?
(Only STEC is more expensive than SF, and not as fast as FIO...)
That's
what I thought too - but I was wrong. |
 |
This could be the start of
new enterprise MLC branding wars in which SSD designers and memory makers
battle it out to try and convince you that their own (different) way of
doing enterprise MLC SSDs is better than all the others. ...read the article | | | |
|
... |
|
| Dataram
talks about blurring memory and storage |
Editor:- May 19, 2011 - Dataram's Chief
Technologist Jason
Caulkins has written a new blog -
Memory and Storage Technologies Begin
to Blur in which he says new nv memories may render legacy hard disk
interfaces obsolete by enabling most storage to fit in servers.
Jason
is no stranger to this concept having designed a
PCI SSD - called
the Rocket Drive 10 years ago. But Dataram's current
SSD appliance is very
much a fibre-channel SAN
animal - designed to work alongside
legacy installed RAID
arrays. Jason says in his blog the key enabler to these new architectures will
be software. |
 |
My
own view is
that legacy hard disk DAS and NAS storage interfaces will continue to
inter-operate with PCIe SSDs (and new faster SSD interfaces) even in a totally
solid state world. I think storage in a complex enterprise will always be
heirarchical - because different types of SSD will be optimized for different
functions - and it won't be economic to use a one type fits all deployment. | | | |
|
... |
|
the Problem with
Write IOPS
the "play it again Sam" syndrome |
Editor:- Flash SSD "random
write IOPS" are now similar to "read IOPS" in many of the
fastest SSDs.
So
why are they such a poor predictor of application performance?
And
why are users still buying
RAM SSDs which cost
9x more than SLC? - even when the IOPS specs look similar. |
 |
This article tells you
why the specs got faster - but the applications didn't. And why competing SSDs
with apparently identical benchmark results can perform completely
differently. ...read
the article | | | |
|
|
|
| |
