Questioning the use of private clouds



Since the rise of cloud computing, enterprises have often preferred the use of private clouds, to public clouds.  This was largely due to control and security issues, or, at least the perception that public clouds had these issues.  So, the operating theory became, “If you’re going cloud, go private.”

The problem with private clouds is that you’re trading off one set of rack-mounted hardware in your data center for another set of rack-mounted hardware in your data center.  The costs are about the same, with private clouds providing a bit better utilization.

Public clouds, on the other hand, allow you to leverage hardware and software you don’t own, as well as provision resources at will, as you need them.  Yes, this model is a bit scarier.  However, it provides much better efficiencies than private cloud deployments, and, in most cases, much better performance and scalability.  The other drawback?  You don’t get to hug the server.

Public clouds have matured a great deal in the last few years.  These days, they’re more ready for primetime than most private clouds. In addition, they’re more cost-effective and do a much better job of providing time-to-market advantages, business agility, and scalability.

These advances are reflected in the market.  The trend over the last year has been to move toward more public cloud deployments.  You only need to look at the rapid growth of AWS to understand that this is more than an emerging pattern.  As public cloud usage explodes, and the deployments are largely successful, the questions begin to form around the use of private clouds versus public clouds, and how to take advantage of the efficiencies that “the cloud” is able to bring.

Although a private cloud is a solid and viable architectural option, it’s typically overused.  Enterprises that want to maintain control of their hardware and software see the private cloud as a way to help them kick the cloud can farther down the road.  They are typically not accountable for the efficiencies of these private cloud solutions, thus success is declared.  No one is the wiser.

There are actually some pretty good private cloud solutions.  Eucalyptus provides an AWS work-alike as a private cloud, and there are literally dozens of OpenStacky private cloud solutions out there, such as Cloud Scaling and Red Hat.  Some provide clear public cloud parings, most do not.

The more innovative enterprises are seeing the advantages of public clouds, including production costs, time-to-market, and agility.  In some cases, it’s becoming a strategic advantage that allows the core business to move faster into new markets, and quickly gets business solutions up-and-running.  Businesses are also removed from the limitations that come with the purchase of massive amounts of hardware and software.

I suspect that private clouds are not losing steam as much as public clouds are picking up steam. Having implemented both, I can tell you that private clouds are a sound architectural option, and they are required in some instances. Public clouds are also a legitimate option, and where public clouds fit, they typically provide more value than private clouds. Most enterprises will end up with both, using hybrid or multi-cloud implementations.

So, many in enterprise IT are now asking the looming questions around the use of private clouds.  While the FUD around control and security are still issues, the news about great successes in the public clouds are stripping away most of the technical and physiological issues around public clouds.

As I said, most cloud deployments will leverage some combination of both private and public clouds, or hybrid or multi-clouds.  While it seems like we’re just trading one complex architectural mess for another, in most cases, these kinds of solutions can’t be avoided.  Indeed, they provide the best options for the business, and have the greatest chance for success.

It won’t be uncommon to see several public and private clouds as a part of most enterprise infrastructure and application platforms in just a few years.  The question then will be, which platforms see the most use?  The answer will come down to what’s the path of least resistance, and what’s the fastest path to getting the business solutions into production, such as new and critical applications.

For the most part, the answer will be ‘public clouds,’ unless there is some legal issue that the application and data can’t reside in a public cloud.  We have a tendency to leverage IT resources in event-driven ways, and that has led to the haphazard implementation of systems that made our enterprise architectures overly complex in the past.  Now that cloud resources are available, and on-demand, that will likely be the path that most in enterprise IT will follow.  Some private, but mostly public.