increases in IT efficiency, as server and storage functions in individual corporate data centers (which are highly inefficient, often running nowhere near capacity) could be offloaded to the cloud. That, said Carr, would push down what’s in effect an IT tax on industry and free companies to invest “in innovation, where it really matters.” What’s more, he said, customers are going to get the best of both of the previous worlds—meaning the high efficiency of the mainframe era and “the personalization that has been possible with the PC age.”
But while cloud computing is making computing power extraordinarily cheap and available, there’s still a lot more we don’t know about the transformation than we do know, Carr argued. Already we’re beginning the see different layers of the cloud emerge: infrastructure as a service, platform as a service, software as a service, and desktop as a service. Taken as a whole, we have yet to figure out “how to harness the world wide computer,” Carr said.
Irving Wladawsky-Berger, chairman emeritus of the IBM Academy of Technology, took us back far further than 1851—to the dawn of single-celled organisms and then to the Cambrian explosion, when evolution seemed to have perfected the cell, allowing an incredible diversity of multicellular organisms to emerge. He views cloud computing in this same framework. Now we are at a time when the evolution of the Internet and the power of grid computing are coming together, he says, providing virtual access to all kinds of devices, sensors, and services, with massive scalability. This will help enable real-time information for congestion management, analyzing supply chains, tracking ships, and a lot more, as everything will have digital components. Wladawsky-Berger sees it all as part of the evolution of the Web—which we originally accessed alone and now are accessing more as social groups—into a much more immersive experience.
After the two keynotes we had two panel discussions and a mid-afternoon keynote from Josh Coates. And there were lots of audience questions. Let me hit some additional highlights.
Rich Zippel spoke on Sun’s Project Caroline, an attempt at developing a new hosting platform for cloud services. Zippel charged that some companies (including some represented on the panels) are looking at the cloud as something they can do on top of existing computing systems—like putting an internal combustion engine on a buggy but not bothering to build asphalt roads to run the buggy on, to use his example. Project Caroline, he said, addresses the idea that you need a very new approach to dealing with the cloud. “Our job is to make the rest of [Sun’s] business hell,” he joked.
Zippel also noted the surging demand for cloud services that he sees among Sun customers. “What shocked us was every single large company we have wants cloud technology badly,” he said. That is, they want the advantages of the efficiency, easy deployment of applications, and utility scale that the cloud promises–but they don’t necessarily want to do it outside the company. That echoed a huge concern of the audience members–security. A number of people asked how they could trust cloud providers to protect key data.
Sim Simeonov of Polaris made the point that the cloud brings incredible computing power to startups and small companies of 3-20 people, lowering the cost of experimentation substantially and making it possible to launch a new company with fewer people than before and almost no IT infrastructure. But he also stressed that the cloud can make it possible to launch “new kinds of businesses” that might not have found firm footing before.
It was somewhere in here that Sim remarked, “Ask not just what the cloud can do for you, but what you can do with the cloud.” If you have a company that operates on Facebook and uses Amazon Web Services (AWS) for its hardcore computing, for example, you can easily handle spikes in traffic and only pay for the services you need. Simeonov told of New York-based Animoto, which uses AWS to turn photo albums into movies and experienced a dramatic increase in traffic—from a relative handful to 700,000 users in three days—when the service went viral on
