It's become all too easy to talk about the cloud as if it is a single, overriding technology affecting data center infrastructure. But is it possible that the cloud is actually less than the sum of its parts? That what we are seeing are really multiple, inter-related developments that collectively fall under the cloud paradigm but nevertheless serve different purposes and provide different capabilities?
GlassHouse Technologies' Ken Copas highlighted this theme recently on wired.com, noting that in addition to public, private and hybrid clouds, there are various flavors of service models such as IaaS, PaaS and SaaS, as well as multiple deployment possibilities like enterprise clouds, commodity clouds and the like. All of these options feature distinct attributes that are suited to specific workloads and will integrate into existing data architectures in unique ways. This makes it incumbent on the enterprise to not only get on the cloud, but to understand what kind of cloud they are deploying, what its strong and weak points are, and how well it will meet stated objectives.
If you think none of this matters, that "the cloud" can be configured and reconfigured to suit whatever needs arise, you might want to take a look at a recent white paper from Cisco that delves into some of the more crucial differences between the various cloud models. Cutting capex and opex is always a welcome benefit to any new deployment, but the extent that each type of cloud can accomplish these feats will depend very much on things like how security and compliance are maintained and enforced. It doesn't do any good to cut costs in one budget only to see them rise elsewhere, which is why all deployments, whether cloud-based or not, need to be evaluated on their merits.
Of course, all successful technologies go through rapid introductory periods featuring high degrees of experimentation, followed by acceptance and consensus on the general parameters of their utility and capabilities. However, cloud technology does not seem to follow that pattern. If anything, it is growing increasingly diverse as new platforms and applications make it easier to move data environments away from on-premise infrastructure, according to Australian research firm Telsyte. Expect this trend to accelerate as more knowledge workers adapt to the kind of mobile, collaborative applications that are uniquely suited to cloud platforms.
Clearly, then, the cloud is not a magical answer to all your IT problems. Rather, it is a targeted technology that can be deployed to meet specific objectives. And the more you know what those objectives are and how your existing infrastructure needs to be augmented in order to meet them, the more successful you'll be. Open standards community OASIS recently identified six crucial aspects of the cloud that must be understood before any content is transferred. These include the provider's infrastructure and performance history, the kind of data and applications to be ported over, the users it will serve and the laws that will apply, particularly the privacy and data protection statutes in place.
Despite the ongoing confusion as to what is and is not the cloud, it is nonetheless an apt description of the numerous computing environments under development. Real clouds may consist primarily of water vapor but can nevertheless come in many shapes, sizes and consistencies. Sometimes they are gentle, fluffy giants drifting across a bright blue sky, and sometimes they are dark, ominous creatures that arise out of nowhere and leave devastation in their wakes.
As in real life, it's best to have a full understanding of what you are getting into before you take that first step.