Having understood the cloud computing model, let us graduate ourselves to the next level, where we will get an understanding of the three different service delivery models in cloud, which is to say in simple words, what can I get from a CSP?
To understand the cloud service delivery models, we need to focus on what we need from a cloud platform and what is provided to us by a cloud service provider; together, they define the as-a-service paradigm of the cloud or the cloud service delivery model.
What do we mean by as-a-service? As-a-service is a new business model, where the consumer does not have to buy a product or solution in its entirety for lifetime use. In the as-a-service model, a consumer buys the rights to use a product or solution for a defined period, in its entirety; the period can be renewed or continued based on the consumer's requirement. There is no lock-in or upfront huge investments needed in the as-a-service model. Applied to IT, this as-a-service model is cost viable for most small to mid-size organizations and start-ups. This has allowed for exploitation of software solutions to bring in business transformation through rapidly developed ubiquitous and innovative applications.
Let us take a look at what the cloud service delivery models are. In the broad category, we have three cloud service delivery models; there have been other specific categories as well, but to discuss all service delivery models is outside the scope of this book. The three cloud service delivery models are as follows:
Infrastructure as-a-service (IaaS)
Platform as-a-service (Paas)
Software as-a-service (SaaS)
The following figure shows a pictorial definition of each of the as-a-service models. IT within an enterprise consists of the nine broad layers of infrastructure components, such as networking, storage, servers and virtualization, and middleware; application infrastructure components such as operating system, and runtime; and a software stack of applications and data:
Cloud service delivery models
In a traditional on-premise environment, all the layers are procured, managed, and maintained by the consumer, who is typically an enterprise. If the consumer opts to not invest in infrastructure and instead leverage infrastructure provided as a service by a cloud service provider, then the consumer is operating in an infrastructure as-a-service model; the stack above the infrastructure components are still procured, managed, and maintained by him. This model is useful when there is a need for elasticity in infrastructure to support ever-changing business demands. This model absolves the consumer of being locked down by huge investments in infrastructure if he has to support a shrinkage in his needs. On the other hand, it also helps the consumer expand his IT in response to the increase in business need.
Taking this discussion to the next level, in the third stack from the figure, we see that in addition to the infrastructure components, we also see a model where the cloud service providers provide the application infrastructure components as a service along with the infrastructure components. This model is called platform as-a-service. Essentially, what this translates to is a multi-tenanted environment where development and deployment of applications can be done collaboratively with the support of value-added software products and solutions offered as a service on the same platform. This model is useful when the consumer wants an easy bring up and tear down development environment, where quick and dirty prototypes can be built. The prototypes can be graduated to a full-fledged production-ready application that is hosted and maintained on the same platform. Again, as with any other as-a-service model with PaaS too, the consumer is not vendor locked and is absolved of upfront huge license costs, when he is still evaluating technology, an exception to this would be when the customer brings his own license (BYOL) to use a specific software as-a-service on the PaaS.
Lastly, software as-a-service is a model where the entire application stack from the infrastructure components all the way up to the application and application data are vendor managed. The user is merely an end consumer of the software application.
As we move from a traditional, on-premise environment to a completely vendor managed software stack, the flexibility and customizability options with the consumer decreases; also, the control that the consumer will have on different layers of the stack decreases. However, the lower costs and speed and agility with faster time to value are achieved as we move from the traditional environments through the types of as-a-service models.
