The Data Center of 2009: Leaner, Greener, Denser, More Modular

More than a decade ago, some IBM researchers began pitching a bold vision of information technology, called utility computing at the time, in which processing power would be made available just as electricity is available, as easily accessible as plugging into an outlet in a wall.

That vision has become a reality in 2009, through the growth of the cloud-computing model. However, this is only part of the picture of the dramatic changes shaping the data center of 2009. We have been seeing the " Next Generation Data Center" taking many forms—from green, power-stingy configurations; to modular, mobile centers that can be shipped to environments where they are most needed; to hybrid centers that support both onsite and cloud-based services. But just as relevant is the recasting of core legacy assets, those large systems that still power many infrastructures.

With all the changes, there has been no shortage of predictions about the impending demise of data centers as we know them. One systems integrator CEO even went so far as to say, "We might just be ready to declare the 'time of death' for the enterprise data center." If the enterprise data center is going away, it's definitely not going to happen anytime in 2009, or anytime soon after. If anything, there has been an unrelenting expansion of data centers in recent years, as organizations have sought to increase their online capabilities. Organizations have made and continue to make enormous investments in both their legacy and Web infrastructures.

In fact, an analysis by Microsoft researchers determined that "data centers now being built contain upwards of 100,000 servers, and the increase in the number of servers that need to be interconnected has stretched the limits of enterprise networking solutions so much that current architectures resemble a spectrum of patches and workarounds for protocols that were originally intended for use in enterprise networks orders of magnitude smaller."

If anything, the problem is containing the managing these fast-growing environments. And that also presents the challenge of finding enough of the right people to keep up with the growth of these environments. Unfortunately, many enterprises are putting premiums on the "hot" skills—such as Web 2.0-related development—to the peril of their data center operations.

While there is plenty of nervousness about what impact the turbulent economy will have on IT budgets, a more long-term issue looms—finding the skilled professionals who are well-versed in “legacy” technologies and languages and can help manage the growth of fast-evolving data centers. While there’s some question as to how much companies will invest in new systems and applications over the coming months, there’s no question that there will be renewed efforts to beef up and modernize existing infrastructures, through service-oriented architecture, virtualization, and other strategies. To underscore the gravity of the issue, Micro Focus recently released a survey that concludes organizations are not investing enough in their information technology to make it sustainable. Nearly two-thirds (60%) of the CFOs interviewed across Europe and North America say they see the skills to modernize core IT assets "as the most valuable in a recession."

But in what Micro Focus calls a "ticking time bomb," it found that fewer than one in seven (13%) CFOs were convinced that they had enough skills available in their organizations to maintain their core IT assets into the future. Instead, the survey concludes, there's too much attention being paid to newer Web 2.0 technologies-"at the expense of the crucial skill-sets required to future-proof the core systems that are most business-critical to the successful execution of operations."

Again, the problem isn't shrinking data centers that are going away; it's data centers growing faster than organizations can keep up. New approaches emerging on the market point to a more manageable solution—what I would call the "loosely coupled" data center.

The technology has been rapidly falling into place. In their report on managing data center growth, the Microsoft researchers propose the implementation of programmable commodity switches to drive down the cost of the network while simultaneously increasing performance through the use of load balancing technology. Microsoft, Sun, and other infrastructure vendors have been also pitching mobile data centers—housed in transportable shipping containers—that can be located wherever required.

Add to that the growing infrastructure-as-a-platform cloud-based offerings that provide processing and storage on demand. And, importantly, cloud doesn’t just mean services delivered from an outside provider. It may even be more commonly seen as services delivered within the enterprise, perhaps by IT, to various business units. The Next Generation Data Center will increasingly rely on various strategies, including virtualization, grid and blade computing, to manage and make services available to end-user groups.

A new study by Blade.Org—the industry consortium of server blade vendors—provides more evidence that we are moving to this kind of configuration. That is, a scaled data center environment in which resources are added or removed as needed. The Blade.Org study says we should expect to see a single "digital thoroughfare" take the place of the multitude of networks that compose today's typical enterprise.

For example, as pitched by Blade.Org, blade computing platforms offer advantages because of their modular design, simplified management and greater integration of storage and data networking components. Virtual, integrated switching within the blade chassis replaces cabling. Capacity can be quickly increased, almost switched on as needed; versus adding physical servers and disk storage boxes.

We’ve come a long way since that original IBM vision of utility computing available from a wall socket. We’ve attained that vision—and also are addressing the requirements on the other side of that socket to make these possibilities a reality. But organizations need to recognize that these capabilities won’t magically happen by themselves—it will still take skilled people to keep things running smoothly 24x7.