Last Updated Apr 21, 2009 11:51 AM EDT
As inevitable as natural law comes the eventual realization that the technofad, while possibly useful, is simply another variation on technology whose full realization is years in the future. And, as industry players make big financial and reputational bets, that's exactly where cloud computing is sitting.
The level of hype on cloud computing is heavy, though not surprising if you have any memory for industry trends. MRP, the paperless office, TQM, client-server computing, object-oriented programming, supply chain management, ERP, the Internet -- all have been technical advances during whose introductions vendors queued to prove themselves on the leading edge. All eventually became important tools for corporate computing, but none offered the polished silver bullet they promised. Right now, cloud computing is thoroughly in the introductory stage.
"The first guy in gets to establish the space," says Lynda Stadtmueller, a senior research analyst at Stratecast, a division of Frost & Sullivan. "That's been the assumption."
In the case of cloud computing, however, the space was not first a vacuum. ASP, Software-as-a-Service, and hosted computing are just three names for the idea of having resources at a remote location available for use. In all these variations -- indeed, as far back as mainframe time sharing -- customers would pay for only the computing done and resources actually used. Where cloud computing is supposed to differ is in its scale and the ability to "separate out the applications fully from the hardware, reach in, grab what we need, [and] combine different software applications in a way that each individual user can use it," Stadtmueller says.
And that simply doesn't yet exist, no matter which vendor is presenting its case. For example, Salesforce.com has been trying to position itself as a cloud computing platform, but Stadtmueller takes exception.
"It's primarily one on one, end customers reaching into and using Salesforce.com and their instance of it," she says. "What you're not seeing is the real benefit of ubiquity and scalability, where I could integrate Salesforce.com with other software in the cloud. It could get there, but today I don't see how it's different from software as a service."
"If you look at even the IBM open cloud initiative, they're talking the talk, but they're certainly not walking the walk," says Robert Rose, vice president of marketing and strategy for on-demand web content management vendor CrownPeak. "What is IBM's cloud offering? It's lotus notes and repackaged IBM stuff as a service. All of them have their standards that they're trying to use as marketing weapons, and that angle is certainly not new. Microsoft has been doing that for years. The open cloud initiative is certainly not open. If there is an earnest attempt at that, I think it will come more in the idea of taking a proprietary application and making it portable across platforms. To my mind, I can move it wherever I want, whenever I want without regard to what's running on the back end. I should be able to take my Java app and move it to Microsoft's cloud, and that's never going to happen."
Rose thinks that moving applications and services among different cloud services -- whether those of Amazon, Microsoft, HP, or IBM, as examples -- is likely more practical. But he says that is different from true cloud infrastructure. "Then you start to get into a semantic argument as to whether that's cloud computing. You're storing it in the cloud."
According to Theresa Lanowitz, a former research director and Gartner and founder of Voke Research, each of the vendors has some aspect of cloud computing in which they excel. "Amazon and Google [are synonymous with] big storage and new technologies," she says. "IBM knows how to handle infrastructure. Microsoft is the quintessential software company that knows how to own a platform, development tools, and applications." In addition, Cisco understands the networking implications and telecom carriers are expert in moving and storing the massive amounts of data necessary.
But each of those strengths is just a piece of what is necessary for cloud computing, and to get the full benefit of sharing not just storage, but CPU cycles and lowering the overall cost of computing for companies, will need all of them. For there to be one big cloud and not a number of smaller ones, the industry will need interoperability in the form of standards that no one vendor is in the position to impose. Combine the speed of the standards-setting process with the technical changes still necessary, and big C cloud computing is likely a good five to ten years off.
However, even though CIOs are not quick to trust vendors, they are anxious to lower capital expenses, scale their needs for global organizations, and maintain independence from any one vendor. They are ready for cloud computing, or at least experiments with it. That means vendors will continue the hype run, at least until the Next Big Thing comes along.
"[Cloud computing] is part of our language today," Stadtmueller says. "You'd better stick the word in somewhere or you're going to be at a competitive disadvantage."
Cloud image via Flickr user akakumo, CC 2.0.