What is a cloud computing infrastructure?

Much like an electricity grid shares energy, cloud computing infrastructure shares a set of computing resources over internet connections and provides more or less storage and processing based on consumer demand. This is the characteristic of cloud computing that provides cost savings by avoiding the overhead costs of buying servers and managing them. It also ensures that an organization does not pay for more computing power than they use. For example, if you bought three servers and frequently only needed the computing power of two, you still have to pay for the up-front costs and maintenance of the third server.

Cloud computing is a byproduct of web applications like Gmail and other online storage/synchronization tools. These applications run in your browser much like a native application that’s installed on your own computer, but all of its processing and storage is provided by remote computing resources.

There are five essential characteristics of cloud infrastructure defined by the National Institute of Standards and Technology (NIST). The first is on-demand self-service, which means that a consumer can provision computing resources on their own, as it’s required, without service provider interaction. Cloud infrastructure provides broad network access where capabilities on a network are enabled through standard mechanisms that allow a wide variety of platforms (smartphones, laptops, etc.) to access the capabilities.

Resource pooling is an important characteristic of cloud infrastructure. The cloud provider aggregates computing resources to be used by multiple consumers with changing usage. Different virtual and physical resources are reassigned to different consumers who have little knowledge about the exact location where their computing is taking place. Another characteristic is rapid elasticity, which means that someone who needs to suddenly increase or decrease the resources they are using can do so quickly. This can be essential for websites that suddenly become very popular, or have shifting sales cycles. The final characteristic is measured service, which basically means that cloud systems can measure your resource usage and automatically configure your usage to the optimum level.

Cloud computing infrastructure is normally defined by three areas of service usage: Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-Service (SaaS). IaaS vendors provide the bare computing power that you need to run a website or application. PaaS provides an additional layer of tooling on top of IaaS so you can build and run your application using tools and infrastructure provided by the vendor. SaaS is a completely hosted service, such as an application, that you can buy. You’ll never have to worry about running it on your own infrastructure - the service is built maintained by a third party and runs over the internet.