Data centres power new computing model

Just about everywhere you turn these days, vendors are in the early stages of crafting business strategies around computing as a service customers can tap into over the Web. Much of the reasoning behind this computing-as-utility model is drawn from the early days of the electricity industry.

We have seen the future of business computing and it belongs to Thomas Alva Edison. Just about everywhere you turn these days, vendors are in the early stages of crafting business strategies around computing as a service customers can tap into over the Web. Much of the reasoning behind this computing-as-utility model is drawn from the early days of the electricity industry.

In the 1930s, most major industrial businesses owned electric generators. It later became more efficient to contract out electricity as a service delivered by utility companies because their power plants could deliver power at a much cheaper rate. The modern equivalent of the power plant is the data centre, which is the crucial centre of all information-processing activity for any modern company. The problem is that data centres are expensive to build and maintain. In fact, any startup company that wants to enter a business segment has to figure out how to overcome this barrier to entry.

Fortunately for venture capitalists and investors alike, a bevy of companies such as Exodus Communications and Global Center have emerged to serve as data centres for a vast array of startup companies. These startups are using Web technologies to challenge traditional brick-and-mortar companies that have invested billions of dollars in data centres.

And although it's unproven that this approach to computing as a service will be profitable for companies such as Exodus, the industry as a whole is starting to gravitate toward this model. For example, US startups such as Smart Pipes are offering IP networking services, whereas Hewlett-Packard is gearing up an always-on computing model based on its E-Services strategy. Microsoft, meanwhile, has launched its ambitious .NET strategy, which envisions the creation of software services accessible via the Web. And even IBM, the paragon of data centre computing, is in the process of either building or partnering with other companies to create a global network of more than 50 data centres.

All of these efforts assume that the current move toward ASPs (application service providers) will hold up over the long term. Of course, given the state of affairs in the ASP space, this is a rather big assumption.

The world is not beating a path to the ASP business model because companies in this space are doing a fine job of giving the utility model for business computing a bad name.

ASPs have yet to figure out how to be attractive business partners for the long haul. No single ASP seems able to cost effectively offer the range of applications that a company of any real size is going to require. Most ASPs stress their ability to get you up and running quickly, as opposed to delivering a substantially less-expensive computing model.

The reason for this is that it's almost impossible for ASPs to effectively run multiple applications for different customers on the same hardware. The best they can do is house your applications on different servers. If you have multiple applications being delivered via different services, or for that matter even the same service, you still have to wrestle with all the traditional integration challenges associated with any large-scale application.

It is during this process that most people come to realise the benefits of being close to the applications when it comes to managing and integrating data. But if that doesn't convince you, a trip to the CFO's office should. The US government, through an elaborate process known as depreciation, makes it very attractive to buy capital equipment that can then be written off over as much as seven years. So when you factor that tax benefit into the equation, owning and running your own data centre is still quite attractive.

Eventually, the computer industry will get the technological and economic infrastructure in place to support a true utility model. But until then, remember the sage advice that futurist Paul Saffo delivered during the InfoWorld CTO Forum last May: "Never mistake a clear view for a short distance."

Michael Vizard is editor in chief of InfoWorld. Send him email at michael-vizard@info world.com.

Join the newsletter!

Error: Please check your email address.
Show Comments

Market Place

[]