Infrastructure as a service (IaaS) typically refers to cloud computing, in which a business or individual obtains hardware services on a somewhat dynamic basis. Cloud computing differs from traditional hosting in that the service provider not only owns, operates, and maintains all the hardware, but the user pays on a per use instead of per unit basis. With IaaS, it is typically possible to pay for the infrastructure that is needed only when it is used. By hosting data and associated services in the cloud, businesses can reduce costs and potentially increase productivity.
In a traditional hosting environment, a server or part thereof is typically rented on a monthly or yearly basis. This gives a business or individual access to the resources available to the server, virtual machine, or shared host. Cloud computing provides the same type of service, but typically on a per use basis. Instead of paying for a server, a user pays for specific resources that are used. The bandwidth, storage space, and other resources are typically billed in terms of utility computing, allowing the user to scale easily depending on specific requirements at any given time.
Actual infrastructures that makes up most cloud computing services can be very complex. The many virtual machines that provide the required resources typically exhibit traits that are associated with tier three and four data centers, including reliability, uptime, and available throughput. Since the requirements of an individual client can be distributed throughout the cloud and adjusted up or down as necessary, it is typically easier to achieve better performance than is available through traditional hosting. With a regular dedicated server, there is typically time and expensive involved in upgrading the memory, storage space, or processing power, but these are all factors that can be adjusted dynamically with IaaS.
The original cloud computing concept dates to the 1960s, though much of the groundwork for the current systems was first laid in the late 1990s. Certain companies began to move towards an IaaS type model in order to increase efficiency. In traditional hosting, a company will often maintain substantially more resources than are required for daily use just in case the traffic spikes. Initial development of IaaS was partially in response to this inefficiency, since the model allows the dynamic scaling of hardware as a service. The first cloud computing infrastructure was offered for rental around 2006, and a variety of different companies soon moved into the field.