Paul Maritz: The Future of Cloud Computing

As the computer industry embraces cloud computing, Palo Alto, Calif.-based VMware aims to deliver one of the key pieces of the puzzle. VMware's virtualization software has become a mainstay of most big data centers because it lets engineers allocate computer resources more efficiently. Now the company is pushing what it calls a "vCloud Initiative," to work with partners to build data centers that can support cloud computing. VMware CEO Paul Maritz talked with NEWSWEEK's Dan Lyons about how the next era will unfold. Excerpts:

NEWSWEEEK: What is VMware ' s vision of cloud computing?
Paul Maritz:
You can divide the cloud today into two categories. One is the enterprise cloud, and there is one, for want of a better phrase, that I call the new-age cloud. The enterprise cloud is really about providing the opportunity for existing IT customers to take their existing workloads and have somebody else supply the underlying infrastructure … The other type of cloud is what I call the new-age cloud. This is about supporting fundamentally new applications. It's not about the current applications that are being used in the IT space. Ultimately the two will come together.

What are the biggest reasons for cloud computing to happen?
Businesses are going to want the flexibility to outsource the provisioning of infrastructure to people who can be presumably more efficient at it than they can be. The motivation is going to come really from having other people provide the "plumbing"—power, the day-to-day management, the reliability, uptime and so forth. Businesses will want to have the option of moving their application loads into, and equally importantly back out of, this outsourced infrastructure as they see fit.

From the consumer point of view, ultimately the user wants his information to belong to him and not to any particular device. Increasingly, individuals are characterized by a body of digital information. And that information needs to live on over a period of decades—the rest of our lives—beyond the lifetime of any device you might have. So everybody is going to need somebody to be the custodian of their information. Just like we don't put our money under the mattress anymore, we put it in the bank. So most of us will become customers of an "information bank" and in so doing become dependent on the cloud. You can see this trend already starting with hosted e-mail services like Hotmail, Yahoo, Gmail, etc.

What are the challenges?
How do you operate at this scale, how do you make things essentially bulletproof in terms of reliability, security and privacy. Those are tough things to do. There are a lot of companies who don't realize there is a discipline and an approach to operating a service that is different from developing a software product. Companies have to learn that as they go forward. This is why this world isn't going to leap up into the cloud tomorrow. This is going to be a progression over a number of years that we work our way through.

How does cloud computing fit in the context of the history of computing?
We are in a big transition from a device-centric world to an information-centric world. It's going to be about how do you make the information useful and available and make that the center of people lives instead of specific devices. Devices will have to cleave to the information rather than the other way around. IT infrastructure, the plumbing, will fade away for most users and businesses, and will increasing be left to professional providers.

What will computing look like in 10 years?
You will basically have a body of information that you can invest a lot of time into customizing. There will be a lot of third parties that add value to your information, people who offer you services. Your information will be in a bank somewhere—an information bank—and people will be adding value to it. Whatever device you use, you will expect that device to be able to access that information in the way that you want as opposed to the way the device wants.

Are we essentially going back to that mainframe world?
It's a kind of a synthesis that goes on. It's the old Hegelian thing. You have thesis, antithesis and you have a synthesis. From the centralized but lousy user interfaces on mainframes, we went to the decentralized world of the PC where we've got pretty user interfaces but we have huge fragmentation of information all over the place. Now we will go to a common synthesis that puts the best of both together.

Is there a risk of a lock-in if companies start relying on a cloud provider?
There is going to be the classic tension between the interest of the user who wants things to be standardized, portable and to have choice and the interest of the provider who wants to have a very sticky relationship with the customer. And I'm sure that, just as before, the pendulum will swing one way or the other over time.

Who will run these clouds?
There will be a variety of companies who want to make a business of it. And we don't believe it will come down to two or three guys at the end. We think it will be hundreds of outfits who want to provide these services.

Will there be clouds based on vertical markets? Or will there be some ü ber -cloud that brings everything together?
I think there will be differentiation. For instance: on quality of service. Some people will be able to take on more liability and provide better quality of service. There will be people who will understand a certain industry's needs better than others. I don't think it's going to be a single giant cloud that provides a least-common-denominator service to the entire world.

Editor's Pick