IT practitioners are enamored with the technical elegance of cloud platforms where software virtualization layers abstract physical objects like servers, switch ports and disk volumes into malleable entities that can be divvied up and dolled out to applications and users as needed. Business executives however care little for the technological wizardry and are only interested in how software defined data centers (SDDC) and the cloud can give them a competitive edge.
Part one of my exclusive interview with VMware CEO Pat Gelsinger focused on the former: how VMware has built the foundation for enterprise clouds by drilling into the details of the company’s product strategy. The second half of our discussion focused on the business case and C-level receptivity to the SDDC message.
Selling VMware’s Vision and Embracing Openness
Kurt Marko: You have offered this expansive vision across pretty much every element of infrastructure, and your customers obviously have a lot of heterogenous infrastructures, so how do you see VMware playing with this mixed environment? Are customers receptive to the all-in-one approach that you’re offering? Are they looking for kind of more of an à la carte choice? What’s that discussion like with customers who are thinking about bringing this in on top of what they have?
Pat Gelsinger: Anytime that you say “customers” and you try to give a single answer, it’s hard, so let’s look across that spectrum a little bit and just break it up. When I have the SDDC discussion with a customer, [for example the] CIO of an enterprise account that’s not one of the Fortune 100, but a big account. I cannot end that meeting without having an opportunity for another leg of the SDDC. Just doesn’t happen. Every single one, they’re ready to take a step on one of the legs, but they’re not necessarily ready to take all the legs, right, at that level. Some of those big enterprise accounts have started what we’ll call greenfield projects to say, “In my brownfield environments, over here I’m ready to transform the management. Over here, hey, I’ve got a DMZ issue; I think NSX is the way to address it. Over here, I really just want to deal with branch office storage in a different way.” So across them they’ll have different perspectives that will take a different leg of the SDDC for different projects at that level. But then they’ll often have projects, and we have probably 10-plus accounts that are now doing greenfield projects that said, “I get the vision. I’m not ready to build the total environment.” That is entirely architected new taking all of those legs as well. That’s the exception and not the rule, but we have enough of those accounts where we’re engaging in very, very fundamental ways. They’re saying, “I’m going to build a green environment that truly embodies all legs of the SDDC.”
Now, as we go from those big accounts to other accounts, if we go down to the mid-market accounts and some geographic accounts as well, there you’ll find that there generally is a greater mix of hybrid cloud as well, where some of those are just saying, “Boy I’ve got to move entirely into the cloud.”
And then there are the accounts as well, and you allude to it, where they’re saying, “I do want choice. I am ready.” Things like OpenStack come up as more substantive conversations with those. They may have multicloud environments as well that are important to them.
And maybe the last characterization of customers would be some of the highly regulated customers, where they’re like, “Public cloud? Not in my lifetime.” So there’s a range, right, of conversations.
Q: That leads to a discussion of opening up to the OpenStack world. What is your view of efforts to standardize various layers of this environment, whether it’s orchestration like OpenStack, the networking piece, which more people are looking at with Open Daylight? … You guys are kind of pushing ahead faster than standards organizations can operate. How do you see that playing out in the future?
Pat Gelsinger: Let’s just frame it a little bit, because there are two very divergent views of what OpenStack is. One view of OpenStack is it’s [an] all open source implementation of a cloud stack. The other view, and the one actually written in the website for the charter, is that it’s a framework for building clouds. Now those are two very different perspectives. For instance, let’s take Quantum for networking [a network service API now renamed Neutron]. We’re driving the Quantum API. It’s an open API. And we’re primarily contributing the Open vSwitch [with] API work, etc. And the best embodiment of that is NSX. Now, NSX is not an open source product. But we’re saying, “Hey, we are gonna support this multi-hypervisor set of interfaces, and if you would choose to replace NSX with something else, that’s OK, because those APIs are agreed by the industry.” But we are driving NSX to be the choice into that environment, and we have very high attach rate of NSX into open cloud deployments today as a result. So is it an open source framework or is it an open framework? [emphasis added]
Our strategy is that we are embracing these APIs at multiple levels and we’re going to make them part of our products. So, I will happily tell you why vSphere is the best product by far. It’s the best hypervisor, it has features like DRS [dynamic resource scheduler] and HA [high availability] and so on that aren’t replicated in any of the KVMs or Hyper-V’s or anything else. We are adding support for the [OpenStack compute resource] Nova APIs to vSphere for management and vCAC [vCloud Automation Center] at the provisioning level. We can provision into the OpenStack consumption APIs as well.
We’re embracing [OpenStack] APIs at multiple levels of our products and saying we will increasingly support these open interfaces into our products. And you’ll see us make more announcements on that as we go forward.
The Business Benefits of Enterprise Cloud
Q: Backing up a bit, what do you see as the business benefits of SDDC and this more cloud-like environment. When you’re talking to your peers, maybe not CIOs but other C-level executives, could you talk about what you tell them this allows them to do, whether it’s financial improvements, business opportunities, new services. What resonates with your peer group?
Pat Gelsinger: The big words that are interesting to the CEO-level discussion, or the very senior COO-CIO kind of player; everybody’s interested in cloud just because they don’t want to not be able to answer the question. They may not know what it really means or is associated with, but it’s on their list, but not necessarily because it’s really the business driver. … The things they really care about in this environment are much more: What are my business drivers? How do I get more agility into my business? How can I take the things that I do today and reach customers or present them to customers in new and powerful ways? It almost always pivots around a business discussion, usually around speed, agility, being able to monetize infrastructure and/or data in new ways, or reach customers through mobility and social in new ways [emphasis added]. That’s sort of where those C-level conversations start.
Underneath that, then you’re able to say, … so let’s unravel cloud and let’s unravel software-defined a bit more carefully. Why does that give you more agility? Why do care? Most CEOs look at IT — except when his cell phone stops working or his PC [dies] — view it as a cost line. Then the discussion is: How can those IT investments become business enablers as opposed to costs that affect your bottom line?
When you start talking about some of the data examples, some of the mobility examples, some of the ability to spin up new applications so quickly, or one of the favorites ones I was deeply involved in, V.me, when Visa went to compete with PayPal to start doing mobile credit. It was all about agility, and VMware and EMC were deeply involved in that product, and we essentially built the mobile based-credit environment with Visa in nine months. For a company like Visa, this was stunning speed. [
Other big regulated companies] have those same speed and agility requirements, because they see PayPals and others coming over the horizon, and they say, “I have to have the same speed and agility of public, but given my regulatory requirements, I’ve got to do that all on my side of the firewall. So how can you help me be inside of my firewall being able to have that same cost of operations, speed of operations, scale of operations, do it on a global basis?” That’s how discussion flow with them. It’s really a thrilling time, but it’s also a very frightening time for a lot of them.
Q: I imagine that IT spend plateau that you showed in your keynote also enters into the discussion. “Oh, by the way, I don’t want this to cost an arm and a leg.
Pat Gelsinger: Yeah, particularly in a regulated environment. If anything, IT spend has moderated a little bit. If you just look at it as the average across all businesses, it’s sort of in the 3% to 4% range. What’s happening is that, increasingly the line of business is taking control for some piece of that spend and a lot of that has become shadow IT. And this is actually where a lot of the public cloud stuff has emanated from.
[For example] Here’s a CIO and all of the sudden Fred in marketing, in one of the lines of businesses, is off doing something on Amazon, and his CEO finds out and says, “Well, that was some good innovation. And are we secure when we do that?” And of course the CIO at that point says, “Not on your life.” And so [the question] becomes: How do I reassert some semblance of control but still do it without impeding my ability to have agility?
For us, VMware, that’s where the hybrid cloud story comes in. … Let us build an environment that allows you to set your business needs, wherever they might be. These might shift between lunch and dinner, or [between] internal vs. external, and it might be different in geographic regions as well. Create as much agility or flexibility in where they turn that [public-private] dial.
Q: That brings up another topic, the whole vCloud hybrid. And talk about how that’s doing. You got some accolades here [Best of Interop, Cloud category].
Pat Gelsinger: I was quite delighted to see that. In terms of the global rollout, we expect by the end of the year we’re going to be in all the major markets, either directly or through partnerships: the three key markets in Europe; Japan, China, Australia and North America. If you add those up, you’ve covered 95% of the cloud market. Savvis is our first partnership, but you’ll see us announce several other partnerships this year. We launched U.K. back in February.
[Regarding the customer response] There was a recent analyst report — I was just reading it yesterday or the day before — that named us the number four cloud in terms of enterprise customer interest. I was pretty happy with that since we only GA’d the service in September, so going from nothing to number four in essentially six months, I was pretty happy with that progression. We have hundreds of logos on the platform now, so it’s ramping very quickly. We’ll be announcing new services on the platform as we go through the year that specifically focus on hybrid, being able to tie together the on- and off-premise. So the speed that we’re moving at here is really delightful to see.
Obviously we should have been doing this years ago and we’re moving quickly to get these services in place. … I have this deep sense of urgency to move quickly in this space. But a lot of market conditioning has been occurring, but now customers want real enterprise answers and in that sense, there really isn’t a competitor in the market that’s secure, that managed, that’s compatible, that gives you dynamic flexibility across the two [public and private cloud]. So the market really does feel like it’s ready for this answer right now, and customer response is very positive.
Q: So when you say “dynamic flexibility,” are you anticipating customers run workloads in both [public and private cloud] simultaneously?
Pat Gelsinger: Exactly. In fact, a couple of examples of that. One is Betfair. They run online gaming and betting in U.K. and they have a private cloud infrastructure running a bunch of workload in VCHS [vCloud Hybrid service]. Every time they start up a new betting pool for today’s race or whatever it might be, on a dynamic basis [they] decide whether to run it on-premise or in VCHS. Literally every time they’re start one up, they decide where to put it. So it is truly dynamic, workload by workload.
Another example is Apollo Group. They run Phoenix University. They are dynamically moving workload across the two [public and private clouds]. They want to keep their premise-based data centers full and drive very high utilization, [but] they’re using us for approximately a third of their capacity [by] moving class groups. They’re have lots of people sign up in the fall [requiring extra capacity] that dwindle by spring. So they’re able to do all that in real time on VCHS [vCloud Hybrid service].
Will VMware Build Hardware?
“We are decidedly a software and services company forever” – CEO Pat Gelsinger
Q: I want to talk about the hardware piece. There’s a report about a so-called Project Mystic that you are building, working with your mothership, EMC [EMC owns more than 75% of VMware] on a hardware appliance. Do you see VMware being more tightly involved in hardware appliances and how might that look in the future?
Pat Gelsinger: We will be more opinionated about hardware environments as a natural consequence of the SDDC. Let’s just again think about the history; I mean we had hardware-compatibility lists for servers. We’ve been doing that for a decade. Why? Because we had to run our software on different server configurations [and] we needed to have opinions about it. As we go to the SDDC, we need to have opinions about the network and the storage as well. [For example with the] V-SAN launch [we said] “these are the drives we’ve tested, these are the servers we’ve tested.” So, we do have opinions in the hardware space across all of the data center resources now, somewhat like we had with servers.
The other aspect of it is is that there is an unquestioned shift toward converged infrastructure, [which] I think IDC estimates it somewhere in the $2.5 billion range. Clearly Vblock [Virtual Compute Environment: a cooperative venture between VMware, Cisco and EMC] has been the leader, but there has also been people like Nutanix and SimpliVity, who clearly have started to say, “Hey, it’s not just for big things,” but also for midrange and low-end things as well. And in that, we see that there is opportunity for us to help deliver a more complete, simple-to-use software stack against converged infrastructure.
So for those two reasons, yes we will be more involved in it [hardware]. But we are decidedly a software and services company forever.
It’s a perfect final quote for the interview since, hardware alliances and product rumors aside, VMware is decidedly a software company determined to be as great a force in the cloud era as IBM and Microsoft were in the days of mainframes and PCs.