You have to love the IT industry; it recycles computing concepts more often than clothes designers in Manhattan's garment district. Meet today's latest fashion: cloud computing, You also may know it as application service providers, grid computing, or time-sharing.
The idea behind cloud computing has always seemed appealing. These darn computers are tricky devices, very difficult to configure and manage, and always seeming to have problems at the worst possible times. Who wants to spend their days sifting through network or system logs trying to figure out why an e-commerce server is down? Why would individuals spend their time sifting through oodles of product documentation attempting to configure the company's latest software release so it matches the underlying infrastructure? Let someone else worry about these issues.
That has been the potential appeal of these concepts, which have been around since the Kennedy administration, yet, every one of them has failed and many of them in breathtaking fashion. The implosion of the ASP model during the dot-com bubble was particularly gruesome. One high-profile startup backed by hundreds of millions of dollars had three customers after an extensive (some would say overbearing) 24-month marketing campaign. It was not alone; the misses have certainly outnumbered the hits through the years.
Don't Miss: Running Your Business In The Cloud
One way to tell the success or the failure of a new technology is by its different monikers. If it is going well, it stays. A PC is still a PC today just as it was 30 years ago. If an idea is failing to take hold, clever marketing executives will conjure up new ways to characterize their product or service, and the hype machine is off and running once again.
Consequently, it has to be somewhat disconcerting to cloud computing supporters that once the ASP boom went bust, their initiative has been given so many names. It wasn't that long ago when grid computing was supposed to overtake the industry. Small and midsize businesses would be able to change their data center configurations on a dime, gaining more power, say during the holiday season, and less, say during the summer. Now, software as a service (SaaS) and cloud computing have become the trendy terms to describe the idea of handing over a company's computing infrastructure to a third party.
To be fair, there have been successes among the different initiatives. For instance, Salesforce.com has done quite well. But, overall, the markets haven't become the runaway hits that proponents initially proclaim. So, the question becomes: Why not?
To answer it, simply pick up a mirror and take a close look your company. Do you know of any other firm that operates in the same manner as you? No, because if there was one, chances are one of you would be out of business.
For cloud computing to work, customers must have similar infrastructures and homogeneous needs. A vendor can only make money if it is able to deliver a standard service to a lot of customers. Customizing a service for each client requires time, energy, and money; however, that is usually what customers desire. The companies don't want their operations to be just like the competitor down the street. They yearn for differentiation. This is a key reason why the litany of vendors promoting cloud themes has struggled. They promised customers significant savings, but haven't been able to deliver them because the customization work required often made the services cost prohibitive.
Don't Miss: Paul Korzeniowski's Networking Blog
There also are items, such as inertia, that have hindered acceptance. Why would a business want to go through the process of evaluating and then moving all of its IT resources? Think of how much time, effort, and money is involved in items, such as training.
There also are technical challenges, and they usually start with security. Do you want to explain to the auditors how you "protect" data stationed miles away from your employees? How do you really know that another cloud customer cannot access your information?
There are also performance challenges. With cloud computing, another layer of devices is added to your computer infrastructure. Invariably, that means performance degradations.
The cloud model can work in certain cases, mainly in startup companies with no existing infrastructure or new lines of business for existing companies. But it's not a good fit for most applications, never has been and never will be. Consequently, the cloud computing hype will ebb in the coming months, just as it did with its forebearers. The marketing types will then get busy coming up with yet another name for this infrastructure option.
See more columns by Paul Korzeniowski.
Paul Korzeniowski is a Sudbury, Mass.-based freelance writer who has been writing about networking issues for two decades. His work has appeared in Business 2.0, Entrepreneur, Investor's Business Daily, Newsweek, and InformationWeek.
