Cloud Computing

Cloud computing is a way to make computing capabilities—such as processing, storage, networks, applications and other services—available over a network from a shared pool of configurable resources. Instead of acquiring and operating every computer locally, a user can request resources from a provider and access them remotely.

A widely used National Institute of Standards and Technology (NIST) framework describes five defining characteristics: users can provision resources on demand; services are reachable over networks; providers pool resources among customers; capacity can expand or contract rapidly; and resource use is metered. These features distinguish cloud computing from simply running a computer remotely or storing a file on a network.

NIST groups cloud services into three broad models. Software as a service (SaaS) gives customers access to an application operated by a provider, usually through a browser or program interface. Platform as a service (PaaS) provides an environment for customers to deploy applications while the provider manages the underlying infrastructure. Infrastructure as a service (IaaS) supplies fundamental resources such as processing, storage and networking, allowing customers greater control over operating systems and applications.

The models differ mainly in which parts of the computing stack the provider operates and which remain the customer's responsibility. In practice, products may combine services or draw boundaries differently, so a label alone does not describe every responsibility or guarantee.

Deployment models describe who the cloud environment is intended for and how it is organized. A public cloud is available for general use; a private cloud is reserved for one organization; a community cloud is reserved for organizations with shared concerns; and a hybrid cloud connects two or more distinct cloud infrastructures. Ownership and operation can vary, particularly for private and community clouds.

Cloud services run on data-centre infrastructure. Virtualization and related software can divide physical computing resources into separate virtual environments, while management systems allocate and measure capacity. Resource pooling can make it easier to serve changing demand, although the customer may have limited knowledge of the exact physical location of a resource.

Elastic provisioning can help an organization respond to unpredictable or fluctuating workloads without buying enough equipment to handle every possible peak. Usage-based arrangements may also reduce the need for a large initial investment. These advantages are not automatic: whether cloud use is economical depends on utilization, pricing, data-transfer needs, migration work and the level of management the customer still performs. Moving a workload can also create dependence on a provider's interfaces or make data and applications harder to move elsewhere.

Cloud adoption therefore involves trade-offs among flexibility, control, portability, performance and cost. A service that works well for a variable web workload may not be the best choice for every application or organization.

Storing or processing information on infrastructure operated by another organization raises questions about confidentiality, integrity, availability, data location and access. Security depends on the chosen service and deployment model, technical controls, contracts and how both provider and customer carry out their responsibilities; using a cloud does not by itself make information secure or insecure. Portability and continuity plans also matter if a provider changes its service or becomes unavailable.

In a 2026 review of selected U.S. federal cloud systems, the Government Accountability Office found that agencies varied in implementing key practices, including continuous monitoring and service-level agreements. The review concerned a limited set of systems, but illustrates why organizations must assess controls and service commitments rather than assume that outsourcing removes their oversight obligations.

Cloud computing grew from the broader idea of computing as a utility and from earlier attempts to deliver applications remotely. Encyclopaedia Britannica describes application service providers that emerged in the late 1990s and notes that remote application delivery became more popular under the cloud-computing label about a decade later. The later combination of network access, large data centres, virtualization and on-demand provisioning made it practical to offer computing resources as services at scale. There is no single event that marks the invention of cloud computing; the term covers a set of related technologies and delivery models.