'In a cloud you need to re-structure your IT organization'

Patrick P Gelsinger, the new CEO of VMware, the company that grew from zero to $3.8 bn in 14 years (he was heading EMC earlier, and was Intel's first CTO before that) talks about his priorities, how he will build up the cloud momentum, and what he wants to do with India. Excerpts-

-Patrick P Gelsinger
CEO, VMware

After Paul Maritz, as the new CEO what new will you bring to the table at VMware? Or will you just be pushing your predecessors plans?

It is not that VMware has a problem. Paul has given VMware a solid strategy and the company is in a very good state. So job one for me-smooth transition. Actually Paul has spoken to employees, customers, and partners, and so have I. And everybody is saying, “Very good, let's get back to work”. Also since the company is performing well, so far everything including the transition has been pretty good. It's not like the Marissa Mayer of Yahoo! who has to transform the company. Those are hard situations and they have to take some bold actions. In our case, there are aspects that will require strong focus. We have laid out a very bold vision like the software defined data center, the new company that we have acquired which needs to be operationalized, the 3-layer transition (from PCs to user devices, from office applications to new apps and big data, and from servers to cloud) that needs more refinement. Paul in many respects was a reluctant CEO. He loved strategies, he loved technology, and directions, but he was not an operations guy. VMware is a company of 13,000 people, an operations pervade everything we do. VMware is like a clumsy adolescent that has grown up too fast. As a company it had extraordinary growth but some of the operation disciplines and maturity need more work. That has to be refined, accelerated, and made more mature.

Since you are talking so much about the need, as the new CEO, to focus on ops, what are those specific areas that need special attention?

For instance, we have laid out our vision on software defined data centers. It is a good and bold vision, and we are all excited about it. For that strategy we announced vCloud Suite that requires packaging and delivering all its components including common interface, single-sign on, upgrade policies, and all that is a big engineering task. Similarly, there are so many other offerings with so many different components that have to be packaged together for our partners and customers. We are now trying to convert into single product offering. For all this a lot of engineering and operational discipline will be required.

CIOs are now getting disillusioned with cloud, specially because of its promise to bring down costs. CIOs believe that with cloud you only shift costs from capex to opex. Actually they are saying that cloud is more expensive. How do you plan to convince them?

I would agree with some part of the statement and disagree with some. If you see virtualization as a precursor to cloud, where one server has replaced 4 servers, then we have clearly shown the huge advantage. Opex improvements are far more operationally dependent. If I keep all the people operating in the same way as they were operating in a silo structure, how can they be operationally efficient, because under cloud the structure is changed. Earlier there was a database team that owned database hardware and processes, and HPC team that ran and owned all HPC hardware, the networking team, and the apps team, and so on. They were all working in silos. But in a cloud environment you need to fundamentally re-structure your infrastructure team. Here no one will carry a networking designation, and everybody will be part of this infrastructure team. So you need to re-organize the team and set up new disciplines to take advantage of the cloud architecture. That is why one of the key announcements we have done is Cloud Ops IP that will help organizations in building, operating, staffing, and measuring public and private cloud infrastructure.

This restructuring or re-organizing business of the IT team to really be able to take advantage of cloud seems to be a complex business. Is this easily doable or will it be a tall order?

My answer would be yes and yes. As it is with organizational behavior issues, they are often hard to implement, but once done they are well accepted. We just finished this at EMC. The company is almost 90-95% virtualized. I rode over the IT department there and drove them crazy to get it implemented. Now EMC will start its cloud journey. Earlier 76-78% of the IT budgets was spent on keeping the light on the remaining on news apps and innovation. Today at EMC, just 60% is going to be spent on keeping the light on, and 40% on new apps and innovation. And for all of that has come out of operational savings. This is spectacular, and our goal is to get to 50-50. My point is that if you keep running everything as same and then build a layer of cloud on top of it, you don't get the benefits.

But don't you think this transition will be tough?

While there are these early adopters to cloud-who are moving because they have someone like me beating them up, or a very tech savvy and visionary CIO, or business conditions demand it-there are many who are not sure how to get going. That is why we have announced the Cloud Ops Forum which will guide them. It gives a blueprint for the transition, an org chart for them to move to cloud. It tells them about the skills assessment, and what new skill-sets they need to put in place. And we also give them certification for that. The Forum will tell them about the new practices they will need to put in place, and the support they will require for that.

Why do cloud vendors keep harping about cost savings from this technology, when a more compelling argument in its favor is how they can move a big chunk of their IT budget from regular maintenance to new apps and innovation?

That is the holy grail. I have not met any CEO recently who has said he is going to increase his IT budget in the coming year, keeping in mind the economic scenario. In fact, most of them are saying that if I am spending 4% on IT, I need more value from that spend, and if I can take it to innovation I will be excited.

One of the reasons often sighted for the slow take off of cloud is the absence of large public cloud service providers, because they build economies of scale and offer it to their customers. Will that not continue to be a big hurdle in making cloud popular?

Today there are lots of IT companies that have the scale, it's just about operating more efficiently. If you look at the philosophy of cloud, its all about abstract, pool, and automate. Which means create the level of abstraction, pool all the resources together, and then automate them. On this side, when we are working in silos we are looking at 5-10% efficiency and we want to move to 70-80% with cloud. Those are the kind of numbers we are talking about.

In many of the emerging economies including India, small to lower end of mid-size enterprises often use unlicensed software or pirated software. If they have to move to cloud, they will have to pay. How to you plan to enroll them, or is that market not on your radar at all?

Even pirated software is not free because someone has to run it, and someone has to maintain it. It might be free in terms of its acquisition cost, but it is not free in terms of its operation cost. The beauty of the cloud is that you have to neither acquire it or operate it, you get it as a service. As a technology company the challenge for me is to find ways where innovation is faster and more valuable than free. I have to give more value than what those offering it for free give. Also, if there is an alternative from the open source community that is free, then customers will want that and say since it is free I will pay only for the support. But that is not a compelling business model. Often open source is a collection of diverse pieces, and virtualization and cloud technologies and offerings are hard and complex technologies so offering it for free will not be easy.

VMware had acquired a company called WaveMaker sometime back to get into the PaaS space? What is the news there, and where is PaaS on VMware's radar?

Yes, PaaS is important for us. After WaveMaker we are active with cloud foundry. What we have done with cloud foundry is very unique. It is the only multi-cloud PaaS available today. Cloud foundry is an open PaaS, providing a choice of clouds, developer frameworks, and application services. Still in its beta stage, I think, this is a very bold move that VMware has made with broad industry support. The unique thing about cloud foundry is that it makes it faster and easier to build, test, deploy, and scale applications. It is an open source project, and will be available through a variety of public clouds including Amazon. And next year we will allow it to run on private clouds as well. It is a very strategic asset for us in the space. We already have over 30 partners building on the platform.

As the new CEO of VMware, how will you view the Indian market? Will it be business as usual or will there be something different and new for India?

You see VMware is not strong in BRIC countries, and also not strong in the low end and emerging economies. I have to put the focus on where the money is going. That is very much on my agenda, as I have to keep the company growing.