We're here at the Datacore booth at VMworld 2016, we spoke with you last year.
Can you tell us a little bit about the company and also what you're showing here at VMworld
2016?
Sure, I'm George Texara, CEO and co-founder of Datacore Software.
The company's got over 10,000 customers around the world, 30,000 installations.
Last year and since then we've seen a 250% growth rate in our hyperconverged business
and we're also very proud that our parallel IO has taken off and today we are the record
holder for the world's fastest hyperconverged systems as well as storage.
We're very proud of our software-defined storage as well as the ability to outperform
everybody else in terms of performance and in terms of response time.
Our record performance right now, in fact, is higher than the all-flash arrays that are
out there and even the million-dollar arrays.
Also here at VMworld we're talking about our universal VVOL capability which allows any
storage that's out there to be placed under our software-defined storage and it is certified
as VVOL compatible.
So it's a great way to take existing storage and freshen it up and make it work within
the VMworld environment and we've done tight integration with VMworld so that basically
now you can work as a policy-based system all from the VMworld environment not having
to get into the details of the underlying storage.
Fundamentally, data cores the place for the world's fastest hyperconverged systems able
to do real enterprise workloads, fiber channel capability, high-scuzzy as well as the fastest
storage on the planet.
And do you guys have a demo at your booth that we can take a look at?
Yep.
Excited to show you.
We'll have Jeff Slab give you a demo on our parallel IO and how we achieve the world's
fastest storage and hyperconverged performance.
Sounds great.
Thank you.
So just talking with George and he said do you have a demo of the product that you can
show us?
Yes.
So what we're going to look at today is we're going to look at the latest technology around
data cores parallel IO capabilities within our software.
Today parallel IO technology exists in our product called St. Symphony and St. Symphony
is a storage virtualization product that has been developed over 18 years and now has
culminated in 18 years of advancement over our 10th generation product and what we're
going to show you here in this demonstration is we're going to show you the power of the
impact of what software can have on today's modern Intel X86 based architecture.
So let's go ahead and take a look and see what kind of performance numbers you can expect
with data cores in St. Symphony.
So what I'm going to show you here is a very, very simple test.
Everybody that's been in storage will know this interface, it's the basic IOMETER interface
and the reason why we've chosen IOMETER to do the demonstration is because this is a
very easily accessible and very readily available benchmark tool that's very common and very
well understood in the industry as opposed to the SPC1 benchmark which we recently ran
over the last several months.
That one is a little bit more involved, it's an audited benchmark and you have to be a
part of the SPC community to get your hands on it.
This is something that's readily available so this is why we decided to show this because
this is something that anybody can reproduce in their lab environment today.
So what we're going to do is this is on a Dell server, it's a Dell R720, R730 series
server and as you can see over here on the right we've got a dual socket system, 24 cores,
48 logical processors worth of processing power which is very important for what we
do and then over here on the left with the IOMETER tool we've got a series of workers
pointing at some disks that we have presented from the data core software.
We're going to demonstrate an 8K block, 90% read, 10% write, 100% random workload which
is fairly common in most environments today and we're going to go ahead and click this
off and we're going to see what the power of the processor is able to give us because
data core is a very special storage virtualization software that allows us now to harness the
power of all of the logical processors that are in the system in order to drive IO to
amazing levels at ultra-ultra-low-latencies and that's really the hallmark metric for
determining storage performance.
So let's go ahead and click this off and we'll see very, very quickly we're going to accelerate
past 1.3 million, past 1.4 million, on our way up to 1.5 million.
Now if you notice as we start to reach the pinnacle performance level which should be
around 1.6-ish somewhere in there, the biggest thing I want to point out here is that if
you notice the average response time right now, this is in milliseconds, this is 32 microseconds
of response time for just under 1.6 million IO, so 1.57 million IOPS by 8K block which
is a tremendous performance level for just a straight commodity-based server with Intel
processors that anybody would have in their environment today running any of their applications
on it, whether it be SQL, Oracle, Hyper-V, VMware, all those applications of course
can benefit from this capability.
Now if you look over at the task manager here, you now will see how data core is able to
exploit the parallelism of the processors, we're able to then distribute the workload,
my connection got interrupted there for a minute, okay?
You can see the distribution now of all the IO across those numerous processors that we
have in the system, so again when we talk about data core performance, it's not simply
a matter of talking about RAM as cache, it's talking about being able to rely upon that
RAM driving IO through that RAM system as a caching system and also to be able to exploit
the power of all those parallel processors to be able to get the storage system out of
the way and allow the application to run a full peak performance without having to worry
about the storage being the bottleneck in the stack, so that's our demonstration for today.
Great, well thanks for taking the time to speak with VMBlog.
Great, thank you very much, we're going to start now.
