Yes, I must sadly admit that you are right and we semantic purists will probably lose this one. Plus, in very large companies, where divisions function like companies and a few years back the IT folks would have been talking about “insourcing”, the distinction between a public and private cloud is little more than academic, anyway.

Nevertheless, I remain an iconoclastic evangelist for the true public cloud for one simple reason. As a business federation-on-demand context, it will enable a business revolution where ad hoc aggregations of small agile specialists can form virtual corporations even at the transaction level, a true new new thing. This can go a long way to ameliorating the waste, inertia, and economic imbalance produced by the last century business model, an exclusive zero-sum game played only by giants.

Private clouds may save their owners money; public clouds will make the world a greener, smarter, fairer place.

So too, when the chasm between problem and solution is great, and business users need immediate benefit from an immediate solution they take to the Cloud. The question is by what means. One wouldn’t expect to see rank and file employees jumping on the company jet. Instead they enter a contract with a service provider; the airlines. The Cloud, like the airlines has made a commodity out of something that would otherwise be completely unrealistic simply by function of cost.

Although I doubt that Bruce would invest in any of them, I am quite certain that one can still find those companies that argue their executives are so important or offer such a security risk that they need to have their own jets. Likewise, traditional IT will continue to make a case for their most important applications and data to be in a dedicated DC. But the relative costs will continue to rise and more importantly, trying to remain at a “Constant state of state of the art” will bankrupt all but the largest IT organizations. One thing is absolute, most departmental, casual and situational applications belong in the Cloud, the same way most employees belong on an airliner.

The reality is that despite the generally oxymoron nature of the term private cloud, it is happening right now. As an example, GoGrid, a infrastructure service provider with a application agnostic platform works with bigger brand name clients to setup a private cloud for the client (i.e. still owned and operated by GoGrid, but dedicated completely – including IP address, and VMs without any shared resources) to the client. Some large companies are contemplating setting up dedicated business units to manage their pooled data center resources to manage peak load requirements. Larger companies with mission critical data (e.g. pharma, financial firms) will gravitate towards an internal/private cloud to alleviate security and control concerns, while still being able to take advantage of the multi/re-use tenets of cloud computing. Smaller companies with less issues around security for mission critical data/apps will lean more towards a pure external cloud.

As you quite correctly point out, cloud computing transfers responsibility for (control of) delivering IT resources from the consuming company, the user, to the service provider. This is doubly problematic for most IT people. First, they think of themselves as providers, not users, and second, as we saw when so many of them fought desktop PCs and departmental servers in the olden days, they like to be in complete control. Saying that cloud computing is just a different mode of resource procurement or a different data center architecture is more politic, but it is also untrue.

My bigger beef with Mr. Berry, though, is that he promulgates the oxymoronic idea of the “private cloud”. If cloud computing finds a fair and faithful synonym in “utility computing”, then there is no such thing as a private cloud. It is like calling a PBX (remember those?) a “private phone company”. They are similar in function – they both route calls – but one is a private facility and the other a public utility. A corporate data center that employs SOA, XaaS, and virtualization, whether on premises or outsourced, utilizes the same technology as a computing cloud, but that doesn’t make them equivalent.

What makes a cloud a cloud is that it is “out there”; like power, water, TV, dial tone, trash collection, and sewer lines, it is a transparent, ubiquitous public utility service whose customers don’t need to know or care about its internal structure and function or the location of its physical facilities. Amazon EC2 and Force.com are clouds. Rackspace and whatever Pfizer is doing with SOA are not.

On the supply side, traditional server licenseware companies like Oracle hate the cloud because it kills their profit margin AND their business model. On the consumption side, business leaders love the cloud for the same reasons – it lowers costs and limits lock-ins. IT people are caught in the middle and, as you suggest, at the crossroads of added value and irrelevance. Unfortunately, allowing them to call a virtual data center a cloud just keeps them standing in the crossroads. Our old friend Marc Benioff is right – let’s get on with it.

Mark

Thanks for your comments and astute observations. I think the models for leasing or renting a taxi fit better depending on the perspective you are considering.

From a service provider’s perspective, they would rather be in the “leasing” business where they can add more services as upsells to the original resource (which eventually becomes a commodity), but currently, they are stuck in the “taxi model” where they only provide the resource (i.e compute or storage) and adding extra services would have a higher bar. However, from a end user perspective, a taxi model does indeed fit better.

Turns out that as an after-thought to this article, I was considering how the analogy would hold up with a Zipcar and/or taxi type of a model and you read my mind on this one.

Best,
Savinay