Why Private Clouds Make Sense for IT

My colleague George V. Hulme has a post today over on Cloud Commons with the amusing title, Private Cloud Haters – The Game is on Like Donkey Kong. While the title may be fun, the subject matter is very serious and private clouds actually make a lot of sense — just as much as public ones do — and for similar reasons.

In his post, Hulme waxed about why there are so many cloud haters, then described the way game maker Zynga mixes the private and public cloud.

“Zynga’s goal, as stated in GigaOM was to launch new games into the public cloud, where workloads may be spiky and were also unpredictable. They’d then bring those workloads to their private cloud as demand became known and stabilized,” Hulme wrote.

And there’s nothing wrong with that approach. It’s a perfectly valid strategy, but it’s not the only way to see private clouds, which could also provide a way for you as IT to build your own private set of IT services housed in your own data center — not unlike the kinds of services you may pay a fee for on Amazon Web Services.

The way it works is that you build a services portal with a set of pre-defined services. You can use physical and virtual hardware as you wish, and this will enable you to scale up and down as needed to deal with your own peak demand periods.

But the portal itself could offer storage, memory, software licenses, pre-designed popular items such as a LAMP stack or other enterprise software development tools — or whatever services you require in your organization.

In fact, it wouldn’t operate very differently from Amazon Web Services or RackSpace, but instead of paying these companies, your internal users would pay you. This produces a couple of positive results.

From an end user perspective they buy what they need and no more because it’s costing the department or project money and they probably don’t have unlimited budgets. Users and managers alike also begin to see IT as a cost center and might actually start to see a value associated with IT services, rather than just the service itself. That’s a direct result from seeing what these services cost.

In a client-server system, users might put up SharePoint site for a project, populate it with data and store bunches of document and abandon it when the project is over. There is no real motivation to take the project down because it doesn’t cost any extra to just leave it there.

But if there’s a cost associated with keeping it up, chances are the project manager won’t keep the site live one day longer than necessary and return those resources back to IT.

You can use private clouds as a place you bring services in-house after testing them on public cloud vendors, as Hulme described, or you can run your own set of cloud services in-house, acting as though you were a vendor. But either way, as Hulme pointed out, you can’t just dismiss private clouds out of hand.

They just makes so much sense for IT on many different levels.