Everyone wants to talk about "the cloud" right now. For technology industry watchers, cloud computing is a cocktail party conversation term du jour. For IT leaders, though, it's time to be careful: This whole discussion about cloud computing gets tricky, because cloud computing means different things to different people. And when your colleague or boss drops an article about "cloud computing" on your desk and says, "What's in it for us?," you'd better understand what to tell him.
When companies like Salesforce.com talk about cloud computing, they're talking about using applications via "the cloud," which means via the Internet instead of via the traditional enterprise software model. So when Salesforce.com talks about "the cloud" they're of course also talking about Software-as-a-Service (SaaS), also known as "on demand" software. Different names, same idea. But Salesforce is taking the idea a step further now, with its Force.com platform, essentially a platform that's accessible on-demand to help other software developers create and distribute applications. They call it "platform as a service." As ZDNet blogger Dan Farber and others have noted, Salesforce.com CEO Marc Benioff and Web whiz kid Marc Andreessen were waxing rhapsodic this week about the future of software development via the cloud.
When Google talks about the cloud, they're talking about banding huge numbers of computers together (something that Google knows a little about) to let researchers in those frighteningly computing-intensive fields like medical and security research do massive calculations. Sounds a lot like "grid computing" doesn’t it? The basic idea tackles the same problem, but Google's vision will band together much more raw computing power. At its simplest, you can think of the cloud like this: Many, many multicore PCs and servers working together over high-bandwidth networks.
IBM's got a similar idea, for an offering they call "Blue Cloud," and they want to help customers use huge amounts of combined computing power via the Internet to slice and dice their massive data sets better and faster. It's mostly talk right now, but look for IBM to dial up the volume on this idea later this year.
When Amazon talks about the cloud, they are talking about their Elastic Compute Cloud (EC2) service. When I first heard Jeff Bezos talk about this, at an MIT Emerging Technology conference in late 2006, I thought the idea was simply brilliant. Need extra computing power? Just tap into Amazon's huge data centers, overflowing with compute power. Buy only what you need. Get server access on demand, for whatever project you like. Amazon makes extra money from infrastructure that they have to have on hand anyway to support their mind-boggling e-commerce operations.
Great idea. Made perfect sense. In an era when many of us were already comfortable using applications via the Web, the time seemed right. The tech industry types got the idea, too: For example, Red Hat used Amazon's EC2 to share the public beta of its Enterprise Linux platform in fall of 2007.
Funny thing though: It's early 2008 and many CIOs have not gotten their heads around this concept, never mind gotten comfortable with it. Industry analysts love to point to Amazon's offering and chat about what it will mean to IT's future. Think much smaller enterprise data centers. Think tons of data stored in the cloud, ready to be mashed up in many ways on demand by business users. That's the big picture, and it's an interesting one. But there are a lot of pixels that have to be filled in before most IT leaders can picture it in