Cloud Computing Paradigms

Cloud Computing offers three paradigms and three delivery models. These three paradigms offer differing levels of abstraction but all are delivered on a utility-based, pay-as-you-go basis. The three delivery models encompass public data centres, private data centres or a mixture of both.

Infrastructure-as-a-Service: At its most basic level Infrastructure-as-a-Service (IaaS) delivers virtualisation and scaling of servers. Virtualised operating system images can be uploaded to a cloud provider who, in turn, provides placement and execution of these images on physical hardware within their data centres. In this model, application software is intrinsically linked to the underlying operating system. Customers are responsible for building their own software and the management of the underlying operating system on which it executes. This includes patch management of the operating system, but customers are not usually responsible for physical hardware, networking or storage. Payment models vary but are typically based on resource usage, such as compute, storage and ingress and egress transactions.

Platform-as-a-Service: Platform-as-a-Service (PaaS) delivers virtualisation and scaling of abstracted software packages above the level of the operating system. Applications are typically uploaded to a cloud platform, although some models require full development on the cloud platform itself. The vendor provisions pre-baked operating systems images on to hardware within their data centres. The customer’s software is then automatically installed on to these images. With this model there is a decoupling of software from the underlying operating system through an application development platform. Software is programmed against this platform and underlying servers have this platform pre-installed. Customers are responsible for building their own software but there is no requirement to manage the underlying operating system nor the physical hardware, networking or storage. Similar to IaaS, payment models vary but are typically based on resource usage, such as compute, storage and data input and output.

Software-as-a-Service: Software-as-a-Service (SaaS) is perhaps the oldest of the ‘as-a-Service’ terms, and refers to fully managed application software delivered as a service. Customers do not need to upload server images or software packages and, instead, rent access to the software which has been created and is maintained by the cloud provider. Rather than pay on a per server per hour basis, charges are often per user, per month. In this paradigm the benefit is that customers are neither responsible for the software nor the hardware.