The private cloud is no mirage

Strict definitions may deny the possibility of private clouds, but reality reveals a very different picture, says Lori MacVittie

Whatever some experts may claim, there is plenty of evidence to show that private clouds really exist, says Lori MacVittie.

According to the narrow definitions laid down by pundits, many IT people are apparently thinking about, discussing and trying to implement something that does not exist — private cloud computing. There is nothing fictional about the private cloud.

I have long held the position that cloud is a deployment model — an architecture built on the premise that resources will be shared, processes will be orchestrated and the total costs to deliver critical business applications will be reduced. As a deployment model, its location is irrelevant.

There can be more than one way to implement a cloud computing architecture. Organisations have always tweaked reference architectures and 'pure' models of computing to fit their specific technical and business requirements anyway.

Trials and implementation
Apparently organisations agree, because according to a survey F5 Networks recently conducted on the subject, 83 percent of respondents indicated they were working on private clouds. Of that 83 percent, 16 percent said they were involved in a private cloud trial, 22 percent were working on implementation and 45 percent were actually using private clouds.

But it is not just our research that indicates private cloud computing really exists. Research firm Evans Data found in its first cloud development survey that 29.7 percent of the 500 developers questioned are working on applications for private cloud environments, with 19.2 percent expecting that they will be engaged in cloud development in the next 12 months.

In short, many people not only believe that private clouds exist, but that they are also actively participating in their implementation.

One of the arguments against private cloud computing is the perception that resources cannot be effectively shared within an organisation. That perception is important, because the premise is that only by sharing resources can the benefits of cloud be realised.

But organisational boundaries do exist in the enterprise: departments, projects, lines of business, even subsidiaries have traditionally included their hardware in project costs and refused to share resources. By crossing these inter-organisational boundaries, resources can be effectively shared.

Yet another argument, the weaker one, stands firmly on the inclusion of 'accessed via the internet' as a defining attribute for cloud computing. Because users of private clouds are unlikely to access applications hosted in the environment via the internet, they cannot call what they are doing 'cloud computing'.

Interestingly, our survey found that definitions, including 'over the internet', were accepted as accurate for cloud computing, with 35 percent of respondents — the most for any of the six definitions on offer — indicating such a definition of cloud computing was, in their opinion, perfect.

Definition quandary
Yet many of those same people are actively working on the implementation of a cloud computing environment, which does not necessarily involve the delivery of services or applications over the internet. It looks as if we have another quandary on our hands about the definition of 'over the internet', or perhaps it is the definition of the term 'user' that is in question.

Perhaps we should stop focusing on what we call it and instead focus on what it does. As Shakespeare might have said, were he an IT analyst: "A private cloud by any other name would deliver applications just as efficiently." In the end, people do not care what it is called. What they do care about is the results.

Whether or not we appease the pundits and call it something other than private cloud seems a moot point anyway. Organisations are adopting and adapting the model for implementation in their own datacentres, and that trend is a more resounding recommendation for the concept than any pundits' stamp of approval.

Lori MacVittie is responsible for application services education and evangelism at application delivery firm F5 Networks. Her role includes producing technical materials and participating in community-based forums and industry standards organisations. MacVittie has extensive programming experience as an application architect, as well as in network and systems development and administration.