Why definitions of cloud are creating 'false' debates
Debates about private cloud computing stem from differences in cloud definitions rather than the reality of what is being implemented in the enterprise.
Why, when so many have already begun evaluating and even executing on cloud strategies, are there still so many debates about what is and isn't "cloud"? When we've seen a growing number of stories about enterprises successfully consuming both public- and private-cloud infrastructure services, why are there still so many debates about whether one or the other is a smart thing to do?
The answer, I believe, stems from a growing division (or misunderstanding?) among technology and business decision makers about the very nature of cloud computing. I hesitate to go into cloud definitions, but I believe that people are arriving at one of two conclusions about cloud, namely that cloud is either a business model or an operations model.
What the nature of these models are, and the differences between them, explains so much about why we are seeing differing points of view even within enterprises themselves. To understand, let's explore each one:
Cloud as a business model. The concept that cloud is a way of selling IT capability as services is probably as old as the introduction of Amazon's pioneering S3 and EC2 services. Many in the industry immediately saw the utility of these services, and quickly went on to declare that cloud is something you buy over the Internet on a pay-per-use basis.
You can see this in older versions of the Wikipedia page on cloud computing.
The idea that cloud is about acquiring IT resources over the Internet is also the prevalent view of the consumer market. Storing your photographs "in the cloud" means acquiring a photo service from a vendor over the Internet.
Cloud as an operations model. The other way to look at cloud computing is as a way to operate IT capabilities. I wrote about this in depth some time back, but the core idea is that cloud is not a new computing technology, but it is simply a new way of operating those technologies as on-demand, self-service elastic services.
An enterprise that operates a service that meets this criteria, then, is providing a cloud service. It may or may not be as efficient or cost-effective as someone else's service, but it is a cloud under this model, nonetheless.
You can immediately see the disparity between the two models. If you are looking at cloud as a business model, there is no way that an enterprise can meet the bill unless they use an internal cross-charging model, and even then the cloud "business" will struggle to reach commercial economies of scale.
However, if you look at cloud as an operations model, the value of running an efficient resource pool with reduced bureaucracy is highly compelling, even if you can't reach the efficiencies of a larger public-cloud provider. Given the complexities of moving data, applications, processes and everything else IT to the public cloud, an internal cloud service becomes a highly compelling option.
Given that business decision makers are probably most aware of the business model experience of consumers, and many IT operators are most comfortable with the operations model view, you can see why even internal factions have trouble seeing eye to eye on what cloud adoption means.
On Monday, I had the honor of participating in a panel at the Enterprise Cloud Summit held in conjunction with Interop 2011. The topic of the panel was "The False Cloud Debate," and my fellow panelists represented all sides of the debate quite well. "False cloud" is a term applied to private cloud by some executives of purely public-cloud companies.
Included on the panel were James Watters of VMware, John Keagy of GoGrid, and Peter Coffee of Salesforce.com. The panel was moderated by David Linthicum of Blue Mountain Labs. Unfortunately, the panel wasn't recorded, but yesterday Linthicum provided a decent overview of the conversation.
In short, while the panel disagreed to the degree in which private cloud makes sense to deploy, all agreed that a hybrid cloud model made sense for most large enterprises. One key reason for this? Data has mass, as my friend David McCrory analogized, and it is much more difficult to move large volumes of data to the cloud than it is to "bring the cloud" to where the data already sits.
Right or wrong, the enterprise is moving forward with projects that target on-demand, self-service elastic infrastructure with some form of cost "show-back" to regulate use; services that fulfill a cloud operations model. They are also consuming plenty of cloud-computing services online in a cloud business model.
The argument that private cloud is a "false cloud" is therefore irrelevant.