How Cloud Computing Works: The NIST Definition Explained
The cloud is rented computing delivered over a network. NIST's definition names five traits and three service types, which this guide explains in plain terms.
In this article
"The cloud" means computing power, storage and software that you use over the internet instead of owning and running the machines yourself. A widely used definition from the US National Institute of Standards and Technology (NIST) breaks it into five traits and three service types. This article explains them in plain terms.
The NIST definition
NIST published its definition in September 2011 as Special Publication 800-145, written by Peter Mell and Timothy Grance. It describes cloud computing as a model for convenient, on-demand network access to a shared pool of configurable computing resources, such as networks, servers, storage, applications and services, which can be provisioned and released with little management effort. The document says the model has five essential characteristics, three service models and four deployment models. The ISO standard definition quoted on Wikipedia is similar: network access to a scalable, elastic pool of shareable resources, provisioned on demand.
The five characteristics
- On-demand self-service: a customer can set up things like server time or storage automatically, without talking to a person at the provider.
- Broad network access: services are reached over the network from many kinds of devices, from phones to workstations.
- Resource pooling: the provider's hardware serves many customers at once, with resources assigned and reassigned as demand changes. Customers usually do not know the exact location, though they may choose a country or data centre region.
- Rapid elasticity: capacity can be added or released quickly, sometimes automatically, so it often seems unlimited to the customer.
- Measured service: use is metered, usually by storage, processing, bandwidth or active users, and reported to both sides, which is the basis for pay-per-use pricing.
Virtualisation: the enabling technology
Wikipedia describes virtualisation as the main technology that makes the cloud possible. It splits one physical machine into several virtual machines, so one server can serve many customers and be used more fully. This is how resource pooling works in practice.
Three service models
The models differ in how much the customer manages.
- Infrastructure as a Service (IaaS): the customer rents basic resources such as processing, storage and networks and can run any software, including an operating system. The customer manages the operating system and applications, not the physical hardware.
- Platform as a Service (PaaS): the customer deploys their own applications onto tools and languages the provider supports. The provider manages the servers and operating system; the customer controls the application and its settings.
- Software as a Service (SaaS): the customer uses the provider's finished application, often in a web browser, such as web-based email. The customer manages almost nothing beyond limited settings.
Four deployment models
Wikipedia, which lists these in more detail than NIST's short summary, describes private clouds (run for one organisation), public clouds (offered over the internet to many), hybrid clouds (a connected mix of private and public) and community clouds (shared by organisations with common concerns such as security rules). It also names multi-cloud, meaning public services from two or more providers.
A short history
Wikipedia traces the idea to the time-sharing systems of the 1960s, where many users shared one expensive computer. The word "cloud" was used in 1994 by General Magic to describe the virtual places mobile programs could travel through, and Compaq's 1996 business plan popularised the term cloud computing. Wikipedia says Amazon Web Services launched between 2002 and 2006, with S3 and EC2 arriving in 2006. Microsoft launched Azure in 2010, the same year NASA and Rackspace started the OpenStack project.
What it means in practice
Because customers pay for what they use and can scale up or down, a small team can rent computing capacity that once required buying servers. The trade-off is reliance on a provider: the customer does not control the physical hardware and depends on the provider's service and pricing. Whether that trade suits a given use depends on the situation.
Sources
Share
Keep reading
Related articles
Why Lithium-Ion Batteries Lose Capacity Over Time
Heat, full charges and deep discharges all wear a lithium-ion battery. Figures from Battery University and Apple show what matters most.
How End-to-End Encryption Keeps Messages Private
End-to-end encryption means only the sender and receiver can read a message, not even the service carrying it. Here is how keys, ratchets and safety numbers fit together, and what it cannot protect.
How Video Streaming Adapts Quality to Your Connection
Why a video drops from sharp to blurry and back: how adaptive bitrate streaming cuts video into short segments and picks a quality level for each one.