Q&A: VMware CEO Maritz Outlines Path To Cloud

VMware CEO Paul Maritz has had a good run since taking the helm at the Palo Alto, Calif.-based company three years ago this month. But much of his company's success during that time frame has been tied to virtualization software, as opposed to cloud computing, where the Palo Alto, Calif.-based company sees the future taking shape.

Like many IT vendors, VMware and Maritz still have work to do in getting channel partners on board with the cloud and all of its attendant business model changes. VMware's major cloud infrastructure stack update on July 12 offers the clearest picture to date of how VMware intends to realize its cloud ambitions and get partners to adopt the service provider mentality, both to their customers and within their own organizations.

CRN sat down with Maritz recently for a wide ranging discussion that included details on the cloud stack update, insight into several recent VMware acquisitions, an unvarnished take on the company's relationship with Microsoft, and how it's trying to get cloud developers into its camp.

Automation has been an overarching theme of VMware's product development in recent years. Can you talk about areas in which this has been baked into the recent cloud infrastructure stack updates and the improvements to vSphere, vCenter, vShield and vCloud Director?

VMware started off as a narrow hypervisor company that allowed you to run multiple copies of an operating system on a single physical machine. Increasingly, most of our features are about how to take a collection of machines and get them behaving as a sort of single giant machine, for efficiency and reliability and availability reasons.

One of the features we have is Distributed Resource Scheduler, where we will actually move applications around in the pool of machines automatically in order to optimize performance, or for high availability, reliability or recovery reasons.

Customers are no longer using our software for server consolidation; now it's about how they do their computing to get additional reliability, availability and efficiency, and all that stuff's baked into the products. That isn't something you bolt on afterwards.

When we talk about automation, rather than trying to layer management software on top, we're baking it into the platform itself. The vast majority of code for vSphere goes toward these automations, that was true of vSphere 4 and it's even more true with vSphere 5.

With the Springsource, Zimbra, Gemfire and RabbitMQ acquisitions, VMware has been building out its application stack and development platforms. And with Cloud Foundry platform-as-a-service, VMware is now trying to win the hearts and minds of application developers. This is a new area for VMware -- how are you going about getting developers into your camp?

We look at the world in three layers: For the infrastructure that apps sit on top of, the great virtue of vSphere is that it can handle, through virtualization, almost any existing application. What we're also trying to do for new applications is to get those applications to run well in a vSphere environment. We also want to be able to make money by selling new capabilities to developers of those applications.

Specifically, we've targeted people who are writing their apps in these new modern programming frameworks, like Spring, Ruby, Node, etc. Our view is that there's a new generation of developers who will be building a new generation of applications, and we're trying to accommodate them on our platform and also have the business opportunity. Enabling that developer is the second tier of what we’ve been doing with Cloud Foundry.

Third, we're focuses on how existing and future applications will be delivered to the end user in a world where you can't depend on the end user holding a particular device in their hands. There is going to be a lot of heterogeneity among those devices. We need to give our enterprise customers a way to equip their users with capabilities in a device-independent way.

Next: The Strategy Behind Cloud Foundry

Linux came about as a way for developers to get around the proprietary barriers that existed in the mainframe world, and you've suggested that the same sort of workaround will eventually emerge as a workaround for proprietary cloud infrastructure. How does Cloud Foundry aim to address the proprietary issue?

Our view is inevitably someone would do something like Cloud Foundry, so rather than wait for it to happen and have to react to it, we're putting our hat in the ring there and preemptively offering something there. And we're trying to do it in terms that are genuinely open, which is why we're releasing it in open source.

Is open source the main thing that differentiates Cloud Foundry from other PaaS offerings on the market?

Not the open source by itself. We think that whatever solution comes up, it will almost inevitably be open source -- open source is almost a pre-requisite. Once you're over that hurdle, it then becomes about the specific characteristics that [PaaS] layer has to have.

Cloud Foundry is drawing from the folks who we recruited for that team who've worked for Google and other places. We have baked this into it the technical characteristics that we think that kind of layer needs.

[Cloud Foundry was also built with help from Mark Lucovsky, technical director at VMware, and Derek Collison, chief architect of VMware's Cloud Services division, both of whom VMware recruited from Google].

You and Steve Jobs both see the world moving to a post Windows era. These comments get a lot of attention, particularly in light of the public sparring that VMware and Microsoft often engage in. However, VMware and Microsoft are also working closely on a number of fronts -- can you talk about the nature of this work?

Our customers have a lot of Microsoft products, and our products clearly have to work together with theirs in these environments. Both organizations, Microsoft and VMware, are mature enough to know that that has to happen, and neither of us will look good if we're doing things that prevent the interoperability that customers want.

We compete fiercely with Microsoft, but this is not a blood feud. It's genuine competition, realizing that you have to be governed by the fact that customers won't accept not passing information between the two of us.

Microsoft says Hyper-V is making inroads in the SMB market, and that it's good enough for what most organizations need. The "good enough" argument is common in IT these days as vendors focus on cost savings in their marketing. As a company regarded as the 'Cadillac' of virtualization, what’s VMware's view on the 'good enough' argument?

We have a spectrum of price points, and… why accept second best? Is there some virtue in accepting second best? No, there isn't.

We've tiered our products and taken price points up -- and down -- over the last two years to make sure that people who like the leading technology don't have an unnecessary barrier to getting it. By and large, that has worked for us.

Next: Mobile Virtualization and SMB Strategy

VMware is well known in the enterprise, but not as much in SMB. What are you doing to change this perception?

SMB is where we've created products like vSphere Essentials and taken our price points down specifically to target that area. And we've seen pretty dramatic unit growth down there, all of which is going into the SMB market. It's not like Microsoft is going to get zero percent market share, but I think they're probably surprised with how resilient we've been.

VMware's Mobile Virtualization Platform (MVP) is an interesting product that captured a lot of attention at the Mobile World Congress back in February. Can you talk about how virtualization can be used to deliver the one mobile device for work and play?

MVP is still very much an experiment. We're going through trials to see how people react to it. We need to see how users are going to react to this concept of having two phones in one phone, and how enterprises will like it.

It's an example of having to deal with this dilemma our enterprises customers face. On the one hand, their employees are increasingly not going to be comfortable with being told that they can only have one version of a black laptop with a specific version of Windows on it. There's a huge revolt against that.

Businesses can't stop these new consumer driven devices from getting into people's hands. On the other hand, they're still going to be on the hook to make sure that they're operating in a secure and compliant environment, and that their information doesn't get compromised by a hacked version of Angry Birds and transmitted to Turkemistan, or whatever it is.

So there's a real challenge here. You want to allow users to get access to all the great stuff that's coming out of the consumer world and still maintain a secure and compliant environment. It's going to require different approaches, and MVP is an experiment to learn if this is one of the ways we can try and square that circle.

Some handset makers are simply cramming all of the business and consumer features onto a single device without using virtualization. Can this approach work too?

The advantage of virtualization is that it provides an absolute firewall. The problem is, when you’re cramming everything onto one device, and users are installing apps there from unknown sources, it's very hard for the enterprises to be assured that their world isn't being infected by the consumer's personal world. And virtualization is a very strong firewall to wall those two things off. That's why people are interested in it.

Being able to measure all the moving parts in a virtual or cloud environment is a key requirement for many customers. Can you talk about how VMware's recent acquisition of Digital Fuel helps get you to that goal?

We're early on in that journey, but we believe that customers need to become more efficient and better informed both a producers and consumers of IT infrastructure. If you can't measure it, you can't manage it. Digital Fuel is a step in that direction.

Next: Recent Acquisitions And End User Computing

End user computing is a developing area for VMware, and you've recently added some pieces to your collaboration portfolio with the acquisitions of SlideRocket and Socialcast. What's the rationale behind building your expertise in this area?

Both in the end user area and in the developer area, we're trying to shoot ahead and take some risks. We think there are big changes coming in both how applications are developed and how they're provisioned and consumed. And this is the time to try and get ahead, and that's what we're doing with those acquisitions.

Most VMware channel partners aren't really familiar with these companies -- should they be?

We put these acquisitions into two categories: one is about near term direct extensions of what we're doing. We certainly try and make them available [to partners]. VCenter Operations, which is based on an acquisition of a company called Integrien last year, will be certainly something that has very close adjacency to vSphere. We certainly will want the higher end of our channel to be able to sell that product and promote it.

For other products that are further out, we're holding those back until we've got more development around those products. And then we will bring them back both to VMware's sales force as well as our channel.

VMware has tackled many technical barriers in its history, but like any virtualization and cloud vendor, many of the obstacles you're facing today are psychological ones. For example, companies are afraid about storing data in the cloud. Can you talk about how VMware works to overcome these fears?

In virtualization, the good news is that the psychological barriers are largely behind us. At this point in time, virtualization is not a concept that we have to go and evangelize, it's becoming quite accepted. If you believe Gartner's figures, sometime by the end of the year there will be more applications running on virtual servers than on physical servers.

Cloud is still an issue where people still don’t quite know what it's going to mean for their organization, in terms of how to manage IT, structure internal IT departments, etc. This is an issue that we increasingly have to come to grips with as we talk to our customers and explain to them that there are different ways to think about this. My sense is that cloud is where virtualization was four or five years ago. We're going to see the same cycle play through.

Google has been attracting growing scrutiny from government regulatory authorities. As a former Microsoft executive who testified on behalf of the company in its landmark antitrust case, what are your thoughts when you see Google getting this kind of attention?

As my mother would say, Google is big enough, bad enough and ugly enough to take care of themselves.

That's a political process they're getting into, not a technical process or a legal process. I think one of our mistakes at Microsoft was we thought it was a technical/legal issue. It wasn't -- it was a political issue. And I think if we’d realized that earlier we'd have probably saved ourselves a lot of pain along the way.