In 1999, VMware stood alone at the top of the virtualization heap, making software to help consolidate workloads on desktops with its WorkStation product.
VMware entered the server market in 2001, offering ESX Server and GSX Server
for partitioning, consolidating and managing computing resources. Over the
next few years the company added a variety of products to virtualize various
scales of computing, allowing several operating systems to run on one
machine.
The Palo Alto, Calif., company partnered with server giants IBM, HP and
Dell, en route to growing revenues in several consecutive quarters. This
caught the interest of EMC, which was ravenous for software vendors to
solidify its transition from a pure hardware vendor to a purveyor of open
software.
EMC scooped up VMware in 2004 and has spent the past year integrating the company
despite leaving it independent as a subsidiary, reporting to EMC CEO
Joe Tucci each month. Earlier this week, VMware came full circle, issuing WorkStation 5.0 with new features.
VMware President Diane Greene, who safely steered the company through an
economic downturn to record growth, discussed the evolution and the
future of the company in a recent interview with internetnews.com.
Q: How has your role changed since the acquisition of VMware by EMC
closed?
We’ve been growing so rapidly and they have left us independent, so it’s a
different kind of resource from having a board of directors. There are some
things we can take advantage of that maybe we couldn’t before, such as
growing sales overseas more rapidly. But it’s unbelievable to me how much
we’re still building VMware.
Q: VMware has enjoyed a lot of success, growing revenues each quarter for
the last two-plus years. But were there any missteps along the way, or
things that you and the company could have done better?
I wish I could have grown the business faster. But it was difficult
convincing people that an old-fashioned software company was the place to be
during the time the bubble burst. We never for a second thought it wasn’t
going to be the right way to deploy everything. But on the research and
development side, it’s been harder to get certain people, while the business
side started to bring in real talent.
We’ve been such a fortunate company. We had a lot of luck. When we launched
WorkStation in 1999, we wanted to get a lot of people using it as fast as
possible, so we launched it as a way to run Windows applications on Linux.
Some 65,000 to 75,000 people downloaded the beta from our Web site. And when
we decided to do that we didn’t realize Linux was going to be as big as it
is. We didn’t predict that, we got lucky. We didn’t think server
consolidation was going to be the big driving application either. We thought
there would be more value around provisioning and disaster recovery but then
the bubble burst and everyone had to do more with less.
Q: When that bubble burst did VMware feel it?
No. We’ve always had incredible growth. I think part of it’s because you can
add our software incrementally, it’s non-disruptive and it doesn’t cost that
much. So we sailed right through all of that.
Again, I’d say we’ve been
lucky. I think our products were horizontal enough so that even if we got
the driving use of it wrong it didn’t matter. And I think we got lucky on
the EMC acquisition. Who could have predicted we were going to work it out
in such a positive, constructive way?
Q: That buy surprised a lot of people. A lot of people thought if anyone were
going to acquire VMware, it might be an IBM or HP, one of the companies that
employed the technology to its own systems.
Well, it wouldn’t have been good to have been proprietary to the hardware.
There’s a huge value proposition around hardware independence. We aim to run
on any hardware and any operating system. So you wouldn’t want to be
with one of the vendors when you’re trying to run any.
Getting
acquired by an operating system company or a hardware company wouldn’t have
been the right thing. It would have undermined our value. Now, I did
underestimate how heated the storage competition is [between EMC and rivals
IBM and HP] but then, working with Joe Tucci, we figured out how to navigate
it.
Q: How does VMware view emerging competition, such as Microsoft or
open source startup XenSource?
There’s a competitive landscape starting to emerge where Microsoft is
predictably trying to bundle virtualization with their operating system.
That’s how they’ve always competed. But what they have today is not even in
the same category as our virtual infrastructure. And then you have open
source like Xen [XenSource product] coming in, which is brand new.
Q: Do you think Xen could be to VMware what Linux is to Microsoft? Many
analysts have said they offer the same sort of virtualization proposition,
albeit without the advanced features.
No, because we have a very open approach to things. We give anybody any
interface they want. We have a full SDK [software developer kit]. We’re
adding all of the functionality in the infrastructure. That vague, little
thin piece of virtualization where Xen is headed? We can break that out and
bring the price down for that as well.
Our roadmap has been phenomenal. If
you look at all of the major inventions we’ve brought out steadily since we
launched in 1999, every year it’s a new major thing and it works really
well. How easy is Xen deployed? How well does it run? Does it have all the
functionality you need?
Q: Another would-be competitor, Virtual Iron Software, is claiming to be
more salient going forward because it has created software that can assume
the processing and performance of several connected servers, instead of
VMware’s single machine virtualization. Are you familiar with Virtual
Iron?
I think they want to get bought. They want to make everybody think they’re
VMware, only better. And maybe they will get bought. But in actuality, they
have a very niche market. They claim if you have InfiniBand between your
machines, and you have an application in front of you on a single machine
that needs more resources, you can expand it across machines. So that says
to me that you have to have an application that fully utilizes the whole
server to be valuable.
Well, there is only one category of application that does that, which is
high-performance computing. But most of today’s servers are only 5 percent
utilized. And now multicore is coming out, which makes them even more
irrelevant. And they don’t have the functionality we have. They’re
trying to drum up a lot of hype because at the end of the day, everybody’s
going to figure it out. We looked at doing that but it’s not where the
market’s going. It’s a specialized application, it’s not horizontal.
Q: Can you talk about anything VMware has coming down the pike product-wise?
In the enterprise today, with our software, you can dynamically allocate and
guarantee a certain amount of resources. If you’re not using it, someone else
can use it, but you can dynamically reclaim it if it’s yours.
What we’re
going to be rolling out is distributed resource management, where you will
be able to run your pool of servers as if it is one big server. You will be
able to see what’s happening with the load and allocation. What we’re trying
to do basically is put all the kinds of things you need to do to run the
servers into the virtual infrastructure.
