How Many Cloud Computing Platforms Can we Handle?

Mendel Rosenblum, formerly of VMware, Parker Harris of, Google's Matthew Glotzbach, Amazon's Werner Vogels and David Anderson of UC Berkeley share their vision for the future of cloud computing.

CAMBRIDGE, Mass.-Experts on the issue of cloud computing discussed some of the promises and problems cloud computing brings to the IT industry, including the possibility that cloud computing could result in there being just a handful or fewer major IT solution providers when all is said and done.

As part of an expert panel on cloud computing at the MIT Emerging Technologies Conference here, Werner Vogels, vice president and CTO at, said sometimes he does not even like to use the term cloud because it has come to mean a variety of different things. He said initially Amazon built its "cloud" environment to support its own business and decided to continue to scale and provide resources. "We continue to roll out services, whether you want to call it cloud or not is up to you," he said.

Mendel Rosenblum, former chief scientist and co-founder of VMware, who is a professor at Stanford University, gave his own definition of cloud computing:

""I think the cloud is your software running somewhere other than your own data center and you have to decouple these virtual machines. I think the cloud is just a logical step that I want to run things on somebody else's machines. The virtualization is just a building block for the cloud.""

Meanwhile, Parker Harris, executive vice president of technology at, said Salesforce lives fairly high up on the software stack. However, Salesforce has services it has been building over the last 10 years, and the company learns from each new customer, he said.

Matthew Glotzbach, product management director at Google, said Google "made a big leap forward in cloud computing when Gmail launched in 2004. That provided a really rich, powerful application on the Web ... and it really started to expose the power of the cloud."

Vogels said with its roots in retail, "knows the economics of making money at scale."'s Harris said the companies in the cloud space have to continue to overcome the barriers around getting massive redundancy. "We all look different, but we're all solving the same problem," he said. Moreover, "going too horizontal also has it issues," Harris said. "I think we're still learning."

Meanwhile, David Anderson, research scientist at the University of California, Berkeley, said, "It seems like the economies of scale in terms of hardware and software will reduce us to a handful or a few utility computing providers."

Harris agreed. "I think that's true," he said, "as long as we're building these massive data centers around the world. It's tough to deliver the quality of service and SLAs [service-level agreements] if you can't manage it all yourself."

For his part, Vogels said there will be different players at different levels of the software stack. "Will there be other companies?" he asked. "I hope so. It's not just investment; it's expertise. So it's not just a matter of building the data centers."

Moreover, the cloud enables a whole set of new businesses to exist, Vogels said. "The cloud supports the democratization of business creation."

In addition, some members of the panel indicated that the cloud causes operating system decisions to be rendered moot. "The operating system is going from this thing that's really critical to this thing that gets chosen by the application developer to support their application," Rosenblum said.

Putting it somewhat more simply, Glotzbach said, "With the next generation of computing, the Internet is the platform." This sentiment jibes with Google's introduction of its Chrome Web browser to serve as the foundation for the company's overall platform stack.

"One of the reasons we launched Chrome is because we're pushing the limits of what the browser can do," Glotzbach added.

Meanwhile, Anderson discussed a project he is working on that he calls "fog computing," which attempts to tap the unused power of private PCs and devices owned by people around the world. The goal is to use the underused computing power as a resource pool to tap for scientific research, Anderson said. Already "about 1 million computers are volunteering" in the effort, he said.