Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model promotes availability and is composed of essential characteristics, deployment models, and various service models.
Cloud Definition: National Institute of Standards and Technology (NIST)
There are five key features that the cloud must provide:
- On-demand self-service
- Resource pooling
- Network access
- Scale Up/Down – on demand
- Measured Service (e.g. "Chargebacks")
One of the nice things about the Public Cloud (through vendors such as Amazon Web Services, Rackspace, etc.) is that market demand doesn't lie and it's much more difficult to obfuscate features. Inside an organization, however, you typically only have one choice – what is given to you by your Operations team. Some Operations teams are afraid of the Cloud and will use a bunch of technical nonsense – either intentionally or through ignorance – about how they're employing a Private Cloud.
Before I cover the Private Cloud myths, reset your understanding of "The Cloud" and think of it more as "Utility Computing". Here are the myths that I hear articulated by, seemingly, intelligent people.
Myth #1: "I'm hosting my instance so we're on on the 'Cloud'"
Nope. Simply hosting software or instances on a remote machine isn't Cloud Computing. It's cool, but it doesn't account for an on-demand, self-service process along with other features. Hosting is a part of Cloud Computing, but it's not the only thing.
Myth #2: "We're using VMWare, Xen, etc. so we're using the Cloud"
Nice try. You probably impressed your non-technical peers, huh? Sadly, probably your technical peers as well. Again, virtualization is a key factor in Cloud Computing – particularly for multi-tenancy; however, it's not the only factor. Read the NIST definition again.
Myth #3: "Most of our provisioning is automated, but, of course, our engineers do need to run the automated scripts that provision the environment"
If you aren't able to launch and terminate instances/software on demand within minutes without assistance, you're not using "The Cloud". If a human is required for any of the provisioning (other than filling out a form/clicking a button, etc.), you don't have a private cloud – as it's not on demand. See There's a big difference between automated and automatic
Myth #4: "We've got a cloud, with the exception of metering; It's too difficult to assess on a continuous basis and we like to keep that information private."
If Non-Operations team members have no idea the costs of virtual instances or the purpose of these instances in real time or near real time, (e.g. "Chargebacks"), it’s not a Cloud.
So, actually, the Private Cloud isn't a lie, it's just misused so often in our buzzword-laden industry, that it loses its meaning. The distinction is important because organizations won't realize the benefits of using "The Cloud" if they don't have features such as on-demand self service, scaling the number and size of instances up and down – on demand, resource pooling, etc. Are you in an organization that lies about its use of the Cloud?