VMware CEO Pat Gelsinger heads into VMworld 2015 with the unified hybrid cloud on his mind
It’s an interesting time to be the CEO of VMware.
The company is one of the pioneers of the modern data center with its virtualization management software, which still holds a large market share. Now, it’s trying to convince customers to virtualize their storage and networking, too. What’s more, VMware is building up its public cloud in an effort to square off against giants like Amazon Web Services and Microsoft Azure, and to help customers build out private clouds, it’s also offering converged infrastructure.
At the same time, rumors continue to swirl about what will happen to VMware as part of a possible shakeup to the EMC Federation.
As part of our ongoing IDG Enterprise CEO Interview Series, Chief Content Officer of IDG U.S. Media John Gallant and Network World Senior Editor Brandon Butler sat down with VMware CEO Pat Gelsinger ahead of the company’s annual trade show, VMworld 2015, to discuss all of these issues and more.
Brandon Butler (BB): What is the big message you’re hoping to get across at this year’s VMworld?
Pat Gelsinger (PG): Obviously, we’ve got lots of things that we’re talking about but the highest level is this idea of the unified hybrid cloud platform. It’s a continuation of what we announced earlier in the year with One Cloud – any app, any device. We rolled that out with a big vSphere 6 launch at the beginning of the year, but there’s lots of stuff coming underneath that. New services, new product capabilities, new cloud hosted offerings, new os offerings and networking features. One of my personal favorites is the hybrid networking that I think is huge and differentiated. But there are also many refreshes: a new DR [disaster recovery] service; new object storage services; new versions of NSX; the next generation of OpenStack; just a whole bunch of things to pay that off.
The second biggest message is probably the Cloud-Native announcements that we’ll be rolling out, which are very much about how we build a robust, secure, managed, scalable, lightweight stack for applications that are being solely developed for the cloud, largely in a container-like way. We’ll be doing a critical set of announcements and rollouts in that area as well. Of course we’ll pepper it with a zillion partner announcements, two zillion customer endorsements and many other things that reinforce that broader industry message. Then the meat and potatoes always is hands-on labs, getting certifications and taking our faithful customers to the next level.
John Gallant (JG): Among all that news, what do you see as the single most critical announcement from VMware?
PG: Unified hybrid cloud. The idea of that is so profound. Every single customer has two questions that come up in every single meeting: Tell me how to get to the cloud and how do I deal with security? What we’re describing with unified hybrid cloud is an architected, capable, proven solution that addresses those two needs. Everyone says: I need to make cloud part of my future but I’ve got this enterprise stuff, these 20-year investments in applications, I’ve got cool things going on over here. How do I bring those together? I can’t just forklift and move to the cloud for those things. It doesn’t meet my regulatory requirements. Bringing that together, that’s what we mean by the unified hybrid cloud – a single architected set of services and capabilities that allow you to run on-premises, off-premises and connect the two together in a complete way that also addresses their governance, regulatory requirements and their security requirements.
JG: Last year when we talked, the public cloud portion of your hybrid cloud strategy was relatively new. How are you measuring the inroads you’re making against the Microsofts and the Amazons of the world?
PG: We have thousands of customers on the platform now, so usage is ramping rapidly. Our partner network is in the thousands and they’re using our platform to host their services using our software, so we’re definitely seeing that uptick in momentum. What really excites me is the substantive enterprise use cases that can’t be done by any other cloud. Some of those that we’ll be talking about at the show are really demonstrating the value that they get from connecting their on and off-premises environments, allowing true hybrid workload ability. I can test and develop in the cloud, I can do production deployments internally, I can connect my private data access to a public cloud infrastructure over hybrid networking, I extend my DR footprint, I’m able to offer VDI deployments in a scalable, automated fashion. These are things that really exercise the difference that we’re having.
JG: Is there a single measurement of success that you think is most critical? Is it the number of developers that are building to your platform? Is it the number of workloads that are moving there, the number of providers or partners?
PG: I’d say it’s enterprise customers. Obviously, when you get an enterprise customer using the platform in a differentiated way, you have a customer for life. For instance, a CIO I was on the phone with just last week said I could not do this with Amazon or Microsoft, they could not solve my business problem that way. Those are the ones that to me are really powerful and that momentum is really starting to reinforce and build on itself. These are enterprise customers, enterprise use cases that are starting to ramp in substantive volume using the platform in a hybrid and, increasingly, in a mission-critical way.
BB: You mentioned you’re going to be announcing some services around disaster recovery and object store. But some of the analysts we’ve spoken with say that Amazon Web Services and Microsoft Azure still seem to be leading the market in terms of breadth and depth of features on the public cloud side. Do you feel you need to gain feature parity with them over time to compete in this cloud market?
PG: Our view of that is VMware must be differentiated and we’re going to do that in hybrid and mission-critical capabilities. It’s not a question of are we ‘jacks-or-better’ with what they’re doing? It’s a question of us doing things that can’t be done by anyone else that are mission-critical, enterprise valuable. We need to have a pretty complete set of offerings that are jacks-or-better as well on the public cloud side. But if they have a particular feature that I don’t have, that’s not going to differentiate a mission-critical enterprise customer who says ‘I can’t do my core business processes on any other platform’. That’s why we’re going to win or lose. I do have to make sure I keep the other aspects of the platform current, but we’ve been able to grow rapidly our customer base, our VM accounts. We’ve been able to do that, and I haven’t had an object store. Truly, I should have had an object store sooner in the marketplace. Now I have one that’s enterprise grade, that meets the enterprise requirements, allows you to be both on and off-premises for enterprise object store. It’s differentiated to meet enterprise needs, but that didn’t deter me from being able to grow the business rapidly over the last year.
BB: Are you seeing any customers switching over from Amazon or Microsoft over to VMware?
PG: Yeah, but in a lot of cases I don’t think it’s going to be a one-cloud world. There are going to be a lot of applications, mobile front ends, other things that people will have on an Amazon or a Microsoft. Some people are going to say that works and when they look at VMware they’ll say – those edge applications, they’re fine where they are, I’m not going to port those necessarily. But the core stuff, that’s what we’re out to be highly differentiated around; the stuff that really matters to the business. Perhaps a customer is taking advantage of Office 365. Okay, that’s not mission critical. How are you running the core databases? How are you connecting your data center network with secure, private networking into a set of cloud resources? That matters to how a business is run. That’s the stuff that we’re going to win. Folks will also look at us and say ‘manage all that other stuff as well’. So products like vRealize Automation to us is critical because we do full workload management into heterogeneous clouds. The IT manager can now say ‘You can manage all my workloads? Even onto other clouds?’ Yes. We’re going to be highly differentiated in the stuff that really matters to the core IT operations of the enterprise. Okay, now you’re my strategic partner. That’s what we’re pursuing.
JG: Pat, on the virtual networking front, you have news at the conference and you’ve been very vocal about the importance of NSX in the overall product offering. Why aren’t we further along in enterprise adoption of virtual networks? For all of the industry talk about it, it still seems that for the most part enterprises are not adopting virtual networking.
PG: At our last earnings call, we said 700+ customers have now adopted virtual networking or NSX from us. Why isn’t that thousands-plus at this point? The thing about virtual networking is that you’re changing the network. That’s like religion in some respects. People who’ve been trained in a certain model of networking, a certain way you do protocol connections, you manage routers, switches, etc., this is a religious conversion to many of those people and the network must operate for businesses today to operate. People don’t make changes there quickly. You have to be able to present them a very compelling set of proof points as well as use cases that are dramatic and impactful to how they run their business. Some of our early customers in the space like eBay have really embraced it; they run eBay and PayPal on NSX. It runs everything that you do, anytime you touch that website, every application, every service, every packet is running through that platform at scale. That is profound. The speed of developing and deploying applications and services is mission-critical for them. Enterprise customers tend to be more conservative, even though there are huge value propositions. That’s why we’ve very much focused on delivering very crisp use cases like microsegmentation that says ‘we can change your security profile in a fundamental and architected way’. You don’t need to change your switching, your routing, many of your operational characteristics but you cannot do this in a physical networking sense. Now we can start putting locks on every door. Every virtual machine or every virtual service can now be protected by an ingress/egress distributed firewall. This is just a radical change in how you build and architect networks. We’re now seeing that acceleration of customer adoption because we’ve done that hard work. NSX will almost be two years in the market come October. Would I like to be further along? Absolutely. But I compare it to where ESX was in year two and we’re actually meaningfully ahead. Here was maybe the most radical product that’s come across the face of data centers in the last 20 years, and my run rate and customer adoption rate is faster than ESX was in its year two. We really believe that this is profound. We are tracking to our business goals, even though we’re ready for that explosive piece of the growth curve. We think the NSX 6.2 announcement really has a lot of refinements of the product, and it will be one that enables that acceleration of market adoption.
JG: What is the software-defined network tipping point? People hate to change the network infrastructure. It’s costly, it’s expensive. What’s the tipping point that will make this a broadly adopted technology?
PG: I think the huge angst that people have around their security challenges. That is the pain point that we can uniquely solve at this point. And now we have enough adoption for proof points and enough of an ecosystem of solutions for that. We have Palo Alto Networks, F5, McAfee, CheckPoint, Arista, Juniper, Brocade – the industry has built out that complete set of integrations. The use cases are there, the proof points of customers doing this at scale are there, the industry support for it is at scale. I think the pieces are now all in place and that tipping point is nigh in front of us.
JG: In May we spoke with [Cisco’s then- CEO] John Chambers. It was the last discussion we had with him before he stepped down and we talked about the SDN market. I want to just read what he said. He said: “It’s unfortunate that VMware threw a hurdle into our EMC relationship because the economic payback from their acquisition of Nicira, it’s been a disaster. It’s not producing any revenue. What did they pay, a billion for a very early stage company? Disaster.” How do you respond to that?
PG: Three different responses. One, as I’ve said, we’re meeting our financial goals for the business as we acquired it. We absolutely believe that this will prove to be one of the best acquisitions – measured in terms of strategic as well as financial return – in the history of the company, if not the entire industry. Point two, I’ll say this in the nicest way possible: John [Chambers] wanted this asset. In the relationship between Joe [Tucci, EMC CEO] and John, this was a very emotional point between them. I am personally anxious for and looking for a much more collaborative relationship with Chuck [Robbins, new Cisco CEO] because this was a deeply personal issue between John and Joe that emerged. Trust me, no acquisition of a billion dollars-plus happens without EMC’s participation at VMware. You can look at our corporate bylaws to understand that. Then third, the perspective that I would have is I don’t need John to lose; I don’t need Cisco to lose. Virtual networking runs just great on Cisco gear and customers are asking us to do more to make it work better. Over 50 percent of my deployments of NSX today, 700+ customers, are on Cisco gear, which really just represents their market share in the industry. Customers want us to do more and we will continue to meet our customer’s requests in this area to the full extent that Cisco enables and partners with us to accomplish it.
JG: We did one of the first interviews with Cisco’s new CEO, Chuck Robbins, and discussed the same SDN issues. Do you expect anything to change in your relationship with Cisco with the new leadership over there?
PG: I’m very hopeful that it will and, trust me, I will do everything in my power to accomplish a much more collaborative relationship, but it does take two to tango.
JG: Do you think that they agree with you that Cisco doesn’t have to lose in this market working with you, that it doesn’t need to be viewed as such a competitive situation?
PG: We’ll see how they choose. I can only say how we’re going to approach the situation with them, which we believe is in the strategic interests of both parties. Secondly, it’s in the customer’s best interest from both parties and it’s playing out that way in the industry today. When they are fighting us and customers buy our product, I think they lose. If we work together I think there’s more opportunity for this to be a mutual win in the marketplace, but our focus remains exactly the same independent of that answer. We’re going to deliver a radical breakthrough in how networking is done through the software layer and we’re going to deliver huge value to our customers and we’re going to work intimately with every industry partner who we can to accomplish the best value with for those customers.
BB: Let’s discuss the EMC Federation. There’s been a lot of talk and some rumors about what the best structure is for the EMC Federation. What do you think is the best structure for EMC and VMware and the other companies in the Federation?
PG: My view on this hasn’t changed. In a period of industry tumult, being bigger and more strategic as a whole is a more powerful position for the companies and we believe that, through this period, the Federation structure is the best one for both customers as well as for the companies and employees themselves. As we’re working through that period, we’re very committed to creating more of that value and delivering more of that value to our customers.
BB: Would VMware ever buy out the parent company, EMC?
PG: Such rumors are interesting to read in the press but obviously these are EMC board decisions. They have the votes at that level so it’s not a question that I can even address at that point in time. It’s a question to the EMC board of their plans in this regard. We’ve been very clear: We’re going to execute on our business plan, partner well inside the Federation and create more value for our customers as a result because we believe it’s the right strategic position for us to be in at this phase of the industry’s changes and overall position. The industry is in a dramatic period of change. The whole move to a mobile cloud infrastructure, the move from on- to off-premises, from hardware-delivered value to software-delivered value. These are dramatic changes in the industry and we’re quite committed that having that Federation structure is the best way to navigate through those changes.
JG: A lot of the coverage about this tends to be about stock value and the value to the companies and shareholders, but help our IT readers understand why that loose federation of companies is better than one company that has all of this under one roof.
PG: Last week, [for] a very large European bank, a big customer of EMC, a good customer for VMware, we laid out our vision of taking them on this transformative journey to the unified hybrid cloud, transforming from their internal Balkanization of legacy systems, etc., embracing mobile and delivering a new model of application development in the Pivotal and Cloud Foundry world. Every aspect of what we were doing was mission-critical to the CIO of a very large European bank and he’s picking carefully who his partner is going to be because he believes deeply that he has to take radical steps because banking is going to be as dramatically changed as taxi service is being dramatically changed by Uber. They believe fundamentally they have to take some radical steps to transform their infrastructure to enable new applications, etc. It was every aspect of what we do and they brought it to the table because they’re picking. You can name the companies: Is IBM going to be my partner? Is Oracle going to be my partner? Is it going to be driven by Accenture? Who can I trust to take me through that transitionary phase in the industry? Is that a conversation VMware could have done by ourselves? No. Or Pivotal by itself? No. Clearly, bringing those assets to the table is just magic. On the other hand, we had a different customer last week. They’re a big NetApp shop. They want to work with VMware; they see the criticality of what we’re doing at that layer. They have a huge partnership with an SI and they absolutely say ‘I need the independence of the ecosystem. I bought into this strategy in a different way’. It really is our ability to do both: Be first, best, strategically aligned to take customers on that transformative journey like the big European bank. Then there happened to be a CIO of a major U.S. law firm where they have made other strategic commitments right now and to them the individual elements of the Federation are exactly what they’re buying into. There will not be a single winner. It’s not like boom, IBM is going to lose and EMC is going to win them all. That’s not the way it’s going to play out. None of us can presume that it’s going to be one cloud environment; Amazon, Microsoft entirely loses and we entirely win. It will be this heterogeneous environment where customers want choice, they want an increasing amount of flexibility but they want to know who they can trust to partner with them for the bulk of that strategic journey.
BB: I want to ask you about containers, another hot topic in the industry and it seems like it’s a topic that VMware is embracing even more at VMworld. Why do you think this is an important technology for customers and what would you say to people who believe that it’s better to run containers in non-virtualized environments?
PG: The second one is like getting in a car that has no doors, no seat belts, no roof, etc. and driving down the road at high speeds. You can do that and people do that. They’re called dune buggies at the beach; but it’s not a way that people run their lives. Are you going to secure (containers)? Are you going to manage them? Do you have the fundamental capabilities in place to have this as an enterprise offering? That’s exactly what we do when we bring the VMware assets underneath it. We secure the network, we deliver the storage connections for containers. We’re able to use physical, silicon-based isolation for security and protection. How many times has the Linux kernel been patched for security flaws this year versus how many times has the VM been patched for fundamental security flaws in the last decade? These things are just at different scales and the car goes just as fast. There is no overhead, so putting the seat belts in place, putting doors on the car and a roof over your head, etc., it still goes just as fast. There isn’t any overhead to the VM. We’ve clearly proven the efficacy of the technologies that we’re rolling out in [Project] Bonneville and others that we’ve delivered, so clearly we can accomplish what we would say is containers without compromise. Obviously, I’m a little bit passionate about this subject.
Why is this so important? Because it’s about transforming the way that people develop, deploy, lifecycle manage applications. For that, containers is potentially a very profound step forward for the industry. Remember, application development is like fads. Some of these things, SOAs and Java and so on, have been different phases of app development. We think containers is potentially the next big one and as such, we’re going to embrace it because we think the value proposition is ‘don’t look down, IT, look up to the new applications and the business services that you can enable’. That’s what’s profound and we’re going to take care of securing, managing and deploying the infrastructure associated with it. Every CIO meeting I have is around that discussion. How can we have you spend less time looking down, hooking this router to this SAN and enable you to spend more time looking up to create more business-differentiating application value? That’s what we really see as the opportunity for us in this space. We do think that containers are really important. We’ll announce a full cloud-native aspect because if you think about containers and next-generation application development, what we’re doing is really two things. One is in our unified hybrid cloud, we just say it’s another workload. We’ve taken care of it and you can keep running that big, heavy, Oracle database that we virtualized or that lightweight scale-out container app. But there’s also a road where people are saying ‘I’m not trying to build something that’s capable of both. I really just want to have a very fast lightweight capability.’ That’s what the announcements are about, having something that a SaaS vendor, for instance, who’s building a truly cloud-native environment without needing to build compatibility for the last 20 or 30 years of applications, do something that is clean and pure but without sacrificing the security, the isolation and as we call it, the secure microvisor. We’ve taken everything out that you don’t need to deliver in that optimized environment but we haven’t sacrificed the security, the management, the isolation, the networking connectivity, things that matter.
BB: What are you doing to make sure that these messages are resonating with developers and startups? VMware has strong relationships with IT executives and administrators but developers seem to be ruling the day in the cloud market.
PG: It’s a great question. In our view, there are maybe three different aspects to that question as we’re approaching it. One is our IT customers, they have to go back and defend their perspective to their internal line of business and developers. So we’re arming them because they want to go back and say: ‘Containers, we got it. Here it is, you can spin them up tomorrow. Here it is in the self-service catalog, go for it. We got it done and I met my fiduciary responsibilities of that being secure, managed, meets compliance, etc. associated with it’. That’s just one audience and you’ll see us do more of that based on the announcements [at VMworld]. A second audience for us are the new SaaS-oriented guys, people who are saying ‘I started up on Amazon and that was interesting but now it’s too expensive, I need to bring it under control.’ We are doing things to reach those SaaS customers right in the fundamental way and that’s expanding some of our go-to-market to reach that class. The third is speaking more directly to developers themselves. We’re showing up in all those places, doing marketing programs, etc., for example, doing keynotes at DockerCon. We’re also partnering first and best with Pivotal in that area because that is a huge piece of what they’re doing and they’ve been developing that muscle of reaching to that next-generation developer. Many of the things that you’ll see from us is how we’re essentially partnering underneath what Pivotal is doing to deliver our cloud-native stack in conjunction with the Cloud Foundry efforts and their developer outreach programs.
BB: I want to ask you about hyperconverged infrastructure where you have some announcements with your EVO product family. This seems like a really hot market right now with a lot of people talking about it, including VCE, which is now part of the EMC Federation. It seems like your strategy is, and correct me if I’m wrong, around a software play here. If VMware provides the software and customers need to go out and find a hardware match for that hyperconverged infrastructure, does that not add more complexity to the sale? Isn’t the idea of a hyperconverged infrastructure is that’s an easy, simple product out of the box?
PG: What we’ve laid out as our strategy is to partner with the appliance provider with a robust set of software offerings that we’re enabling. We’re being prescriptive and working closely with partners building appliance offerings through their channels and distributors, as well as working with OEM partners. It is a different strategy than somebody who’s providing the full hardware/software. We’re working with the people making our software available, working closely with them to enable that whole appliance offering in the marketplace. Ultimately, we believe that the breadth and depth of that ecosystem will be beneficial for overall competitiveness in the marketplace. Obviously, VCE and EMC invented the converged infrastructure (CI) category. The Vblock began the entire view of what converged infrastructure is all about and everybody – IDC and Gartner and all the other industry analysts – recognized them as the clear leader in creating that category. Now it’s act two of converged infrastructures, and that is hyperconverged. I think VMware gets credit for beginning that terminology. We see this as a natural evolution from Vblock, now sort of down-marketed to a more appliance offering which isn’t nearly as big as a Vblock might be. This is sort of round two of CI and we believe it’s formative, it’s hot and we will be participating as will EMC, and our other OEM partners. VCE will be a partner with us on this too.
JG: Let’s drill in more on hybrid cloud, since that’s a big component of your strategy. What are your key differentiators in particular against Microsoft, because they talk a lot about hybrid too. When you compare the hybrid cloud strategies, nail down the top three or four ways VMware is different in helping you build that really effective hybrid cloud.
PG: Number one is — where are you starting from? Given our huge leadership position in virtualizing enterprise workloads today, that’s our starting point. You’re running virtual on us today for 80, 90, 95 percent of your workloads. We can connect that to the cloud in a fundamentally compatible way. You don’t have to convert anything, you don’t have to migrate anything, if you already have it in the VMware-based VM, done – we can bring that to the cloud. That is a huge starting point for us and given the relatively minuscule market share of any other alternative in the enterprise, that’s a huge advantage. Again, we’re talking about mission-critical stuff not some little test and dev thing in the corner. In the core of what they do our market share is huge. Second it’s not just the compatible VM, but it’s also the management plane that’s heterogeneous. Most importantly is the secure networking layer that leverages heavily the NSX capability. That is a huge differentiator for us. Nobody else has those kinds of hybrid networking capabilities. Not Microsoft, not Amazon, not anyone else. That’s one where truly I can make the public cloud resources of vCloud Air appear like a secure, compatible extension of a private cloud data center of a customer and they just view it as extending resources to another data center. This one you don’t have to manage but it’s seen in a compatible, absolutely transparent way at the networking level. That’s huge. The last piece of that story is one that we’ll touch on a little bit at VMworld, which is bringing more of the mission-critical capabilities into that offering. It’s very much where the Virtustream technology plays a profound role because we can uniquely deliver mission-critical, high IOPS-sensitive SLA workloads in a cloud fashion. That’s another huge differentiator. The last one, of course, is our 4,000+ partners. We’re seeing others mimic our partner program and they might have five, ten or twenty partners. We have 4,000 partners at this point running our software stack in their cloud and that gives us just geographic reach, different vertical segments, different certified clouds in different portions of the market. I was just with a major customer yesterday and we started working down country by country where they would need a cloud service and I would say ‘we got that one, we got that one. Oh, we’ll have to go to a partner for that one, that one. Right here are the three partners we have in Malaysia. Here are the two partners that we have in Jakarta. Here’s the global footprint that we have in South America’. That’s powerful and we’re just well ahead of anybody else in that domain.
JG: We spoke recently to the CEO of Infor, Charles Phillips, and he talked about putting their SaaS offering on Amazon and saying that in the long run it didn’t really make a tremendous amount of sense for them to be in the business of running the cloud infrastructure. In terms of your strategy, could you have done everything that you want to do in helping customers build hybrid cloud but do it on Amazon’s platform?
PG: There are things that we can do on top of Amazon and we are doing some of those things and I think you’re going to see us continue to do more. The best example of that is vRealize, with which we manage workloads for customers on Amazon. It’s probably the best example that we have today. We might wish [Amazon] didn’t exist but it does and customers are using it. How can we best make that part of the fabric of services that we’re delivering to customers through the magic of our software? We’ll definitely enable and embrace that throughout our product portfolio where we can. At the same time, some of the things that I was describing earlier, such as the secure hybrid networking, I can’t do on Amazon. They don’t have that capability on both ends of the wire and that is a huge value to customers. I just can’t emphasize that enough and we’ll talk about some of the customer examples at VMworld. This is stuff you just can’t do on any other cloud. If they would embrace NSX as their fundamental networking technology, okay, we could do more things but even then, the less we have both ends of the wire, there’s just a lot of limitations. For instance, on security, policies, distributed firewalls, other things like that that brings huge customer value. We will work to make those as available across platforms as we can in the industry but fundamentally those are just hard technology problems that we can solve for our customers through the unified hybrid cloud.
