Cloud computing is a current buzzword and the definitions around the web are as varied as the verity. Here, we attempt to define just one part of the Cloud: Infrastructure as a Service.
Prior to Cloud Computing, there were basically two ways of increasing your physical hardware: on-site and hosted co-location at a data center. On-site infrastructure required large expenditures; for power, cooling, floor space, and the actual hardware and enclosures themselves. Co-location only saved you power, cooling, and floor space costs - you still needed to provide the hardware.
With the advent of Cloud Computing, and associated technologies, all of the expenditures are removed. You don’t need to pay for power, cooling, floor space, or the hardware. The costs incurred with IaaS are for storage of data, bandwidth, and hardware usage. There’s no initial investment, so the risks are extremely low - you only pay for what you use, the same as the way the power company bills you for electricity usage. When you move to a new home, the first company you have to call is the electric company. Pretty much every device you have is plug and play with the power company’s infrastructure, and it is billed to you by your usage; there’s no monthly fee. However, as you use more, the rates increase - and that is where we draw the line in our power company metaphor for cloud computing… Cloud prices actually scale down the more you use them.
A benefit of this paradigm is expandability and the eradication of obsolescence of hardware - you never need to pay to upgrade your infrastructure. When your business grows, and you require more power, you can simply request more hardware from the cloud provider. Providers such as Amazon EC2 have defined “Compute Units” that describe an amount of compute capability; allowing you to quickly determine how many you’ll require based on your current needs. As an example, you may need more processing power during business hours, but only a basic amount during off-peak. IaaS allows you to provision and request more hardware on an “as needed” basis - creating a side effect of reducing costs further compared to on-site or co-located hardware.
The way all of this is accomplished is with the use of virtual machines and hypervisor software. In essence, a server at the provider of your choice - when you request infrastructure - is either partially or full provisioned for your use, and becomes accessible generally within minutes.