The use of cloud infrastructure will undoubtedly increase in the coming year. The question is: By how much? Or, more to the point, how much of the traditional data center infrastructure will transition to the cloud?
To hear some people tell it, just about everything data-related will be on the cloud in very short order, resulting in severe, detrimental consequences up and down the IT vendor chain. Barclays, for example, recently issued its 2014 global technology outlook and concluded that overall IT spending will be held to single-digit growth because of the broad federation of resources that the cloud produces. This deflationary impact is the result of every dollar spent on cloud infrastructure producing a three- or four-dollar reduction in enterprise IT spending. To date, most of the cloud spending has come from Amazon, Google and other behemoths, but the common enterprise is now poised to increase its cloud infrastructure in a big way, producing significant changes in the way both hardware and software platforms are designed, developed and deployed.
At this point, few organizations are talking about a full transition to cloud-based IT, but it is clear that virtually everyone sees the cloud as a significantly transformative development. When you get down to it, however, transformation is difficult, which is why most organizations are using the technology for peripheral applications like email and office productivity rather than those that impact core business functions, according to Comcast’s Ian Gotts. Part of this is due to simple ROI calculations: The more difficult it is to switch to the cloud, the worse the cost/benefit ratio becomes, which means apps that have a long list of established relationships with legacy infrastructure will probably be the last to go.
Of course, once a transformative technology like the cloud gets its hooks into the enterprise, it has a way of taking over. Cloud providers are already devising a wide range of architectures designed to serve key verticals like banking and health care or to meet highly targeted, general-purpose enterprise needs, says Equinix UK Managing Director Russell Poole, and these would appeal to enterprise users more for the problems they solve, rather than the technology they employ. And once both public and private clouds become intertwined within hybrid architectures, the pressure will be on to place ever more critical applications on the leaner, more flexible infrastructure.
Already, trade organizations are turning their attention toward simplifying the cloud transition. The Cloud Ethernet Forum, for one, was created earlier this year by IT professionals struggling to overcome the challenges of integrating disparate cloud architectures with each other and with legacy data infrastructure. So far, the group has targeted five areas in need of improvement if the cloud is to fulfill its promise as the next-generation data environment: virtualization, automation, security, programmability and analytics. Unless and until these functions work seamlessly across multiple cloud architectures, don’t expect cloud nirvana to emerge any time soon.
Those of us who have been following technology developments for the past few decades realize that, with rare exceptions, things hardly ever work out as planned. Initial hopes for the earliest consumer PCs were that families everywhere would use them to balance their finances, organize their schedules, educate their children and bestow bliss and harmony upon the land. Instead, they sat in the corner of the living room collecting dust because no one could figure out how to use them—at least until the Internet came along.
The point is that the cloud is still an unknown quantity in enterprise circles. People are pretty sure they want it, but they have yet to fully understand how to make proper use of it. That knowledge will come with time and experience, and it will probably lead to a wide variety of cloud configurations and implementations—one for each and every enterprise on the planet.