We’ve already established that everything that is called “the cloud” is not, in fact, the cloud. Cloud services are much more available than an old server hosted in a faraway data center. When I talk about cloud hosting, I mean the following:

There is no single hardware point of failure in a true cloud solution.

That does not mean that the application won’t have issues, but the failure of a single server or switch should not mean lights out for all of the applications hosted on that hosted infrastructure.

There is no single infrastructure point of failure in a true cloud solution.

I may be idealistic here. Of course, something could always go wrong, but the odds of that are less when there is an investment in redundancy. A dedicated hosting solution, perhaps you have one in your own back office, would require that you have another server available for every one in use, redundant network connections, redundant cooling facilities, and so on. It gets expensive.

Today’s cloud solutions can be reconfigured in real time.

Do you need more capacity? Temporary horsepower for that product launch? Turn it up, then turn it down, whenever you want it. No tickets needed. No humans in the loop.

Let me enumerate some of the cloud infrastructure providers that I know of – and this is by no means an exhaustive list:

    • Amazon Web Services (AWS) (Amazon’s view of cloud computing can be found here)
    • Microsoft Azure (Microsoft’s virtual datacenter tour can be found here)
    • Google Cloud Platform (there is a great video of their OR datacenter here)
    • IBM
    • Rackspace

If you know of others that should be on this list, please let me know.

Let's get to work.