3 simple truths about the cloud

Chris Bronk is a research fellow at Rice University’s Baker Institute for Public Policy and an adjunct instructor of computer science at Rice.

Information technology's biggest buzz item — excluding the iPad — remains cloud computing. For a couple years now, we have heard industry and government talk about the cloud and how it’s going to change everything. That’s mostly — but by no means entirely — hype.

We’ve been drawing clouds between network nodes to symbolize the Internet’s wherever-the-data-needs-to-go routing pattern for almost 20 years now and have scads of slide decks to prove it. So here are three things worth remembering when someone says, “Hey, gang, let’s take it to the cloud.”

1. There would be no cloud without virtual machines.

Each server provides some sort of service. That’s why we call them servers. Some are busy, some are not so busy, and some are busy only some of the time. When a server gets too busy and fails, that is wasteful. When a server runs at 10 percent of its potential capacity because it has little to do, that, too, is wasteful. The answer to that efficiency problem has been virtualization, in which each server is divided into multiple virtual machines. As demand for a service rises, the data center manager can dial up additional processing capability — what computer scientists call cycles — and then ramp it back down when demand falls. That approach is useful, for instance, on the days that everyone in the organization logs on to the human resources service to review their pay stubs. Cloud providers use that concept but on a much larger scale.

2. The cloud is about the Internet, and the Internet is about Web browsers.

Although the Defense Advanced Research Projects Agency, Defense Department, National Science Foundation, scads of academics, and Al Gore invented the Internet, Marc Andreessen and Eric Bina’s Mosaic, one of the first graphical Web browsers, unlocked the Internet for everyday users. Mosaic was good at rendering text and images into a format people could understand, and browsers have continued to evolve. Modern versions support all sorts of functionality beyond serving Web pages. Java and its more sophisticated cousin, Asynchronous JavaScript and Extensible Markup Language, allow browsers to provide functions we’re accustomed to seeing in applications installed on our computers. Anyone who has used Microsoft’s Outlook Web App will tell you there’s not much more the locally installed Outlook client can do that the Web version can't. Note: If your organization does not use modern browsers (i.e., something more recent than Internet Explorer 6), no cloud for you.

3. The cloud needs pipes.

And now the downside: the tubes. Old-school enterprise computing was all about mainframes and terminals. Post-PC, post-Internet computing challenged the client/server model. Programs ran on our machines, but we might also use an application installed on another machine connected via the network to a server with a database, or we might store a document on another server. The genius of cloud applications is that you don't need to load applications on PCs. Travel to Japan, visit a cyber café, load your Google Docs on some random PC, and go! However, the reality of cloud computing is that if you aren’t connected, the cloud is closed. There are some slick ways to synchronize content off-line, but that means you have to truck around a device with all that off-line content, so you can’t ditch the PC just yet.

Those things, among others, make the cloud go. Like all good marketing campaigns, the cloud is more of a concept than a reality. It just took a while for the marketing guys to work with that cloud thing that’s been on our slide decks for years.