Among all the various pieces of cloud terminology, the one that many people, even tech experts, have trouble with is the concept of a private cloud. I’ve heard people describe Gmail as a private cloud service, because users have private accounts. I’ve heard people insist that the only platforms correctly described as private clouds are those where organizations themselves purchase, deploy, and manage the underlying physical hardware.
The second group are closer to the truth, but far enough from capturing the real meaning of a private cloud that it’s worth devoting some time to explaining in simple terms exactly what a private cloud platform is.
In order to cut through the confusion, I’m going to base my explanation on the definition used by a trusted third-party, namely the National Institute of Standards and Technology and their Definition Of Cloud Computing document.
First, we have to be clear about what the cloud is. NIST names five essential characteristics that a cloud platform usually has: on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service. All of these reflect what we commonly consider to be a cloud platform.
To reduce the chance of confusion, we’re going to concentrate on private clouds within the Infrastructure-as-a-Service model. It makes sense to consider Platform-as-a-Service and Software-as-a-Service from a private cloud perspective too, but since each of those service models usually — although not always — runs on top of an Infrastructure-as-a-Service platform, and derives its status as private or public from the status of the underlying IaaS platform, we’ll stick to thinking about the foundational cloud layer.
Private clouds are a deployment model for computing resources that conforms to the essential characteristics of the cloud we just mentioned. Familiar cloud platforms, like Amazon’s Web Service, provide underlying physical infrastructure — servers and networks — on top of which runs virtualized computing resources. The public are able to deploy virtual infrastructure into the cloud, but they have no knowledge of or control over who else runs resources on the same physical hardware. Physical servers, disks, and networks are likely to be used by many different organizations — that’s what makes it a public cloud. Public clouds are almost exclusively owned and managed by third-party vendors.
In a private cloud, one organization has exclusive use of the physical infrastructure underlying the cloud. There may be multiple users of the cloud platform within that organization, but no other organization can deploy cloud infrastructure onto that hardware. Many organizations prefer private clouds because it gives them more control over the physical infrastructure layer, which in turn provides better accountability and security.
Here is the point at which the confusion tends to arise. Note that so far we’ve said nothing about who owns and manages the underlying physical hardware. It’s a misconception that the organization that owns and manages that hardware has to be the same as the organization using the cloud resources. According to NIST:
It may be owned, managed, and operated by the organization, a third party, or some combination of them, and it may exist on or off premises.
A business may well choose to have an in-house data center running a private cloud, but if only one organization uses the cloud platform, then it can be managed by a third-party vendor and still fall within the definition of a private cloud. Cirrus Hosting’s dedicated private cloud platform is of this nature. Cirrus Hosting manages the physical infrastructure and virtualization layers, but they are for the private use of each organization.