How it works
Cloud computing is a model for accessing computing resources over a network, on demand. Instead of buying and running all the infrastructure locally, an organization can contract processing, storage, networking or software capacity offered by a provider.
The definition from NIST, publication SP 800-145 describes characteristics such as on-demand self-service, network access, resource pooling, rapid elasticity and measured service. These properties help differentiate the cloud from simply hosting a server outside the office.
Services are usually presented in categories. Infrastructure as a Service offers basic components, such as virtual machines and networks. Platform as a Service adds managed environments to develop and run applications. Software as a Service delivers a ready-to-use application over the internet. In practice, each provider can use its own names and boundaries, and some products mix these categories.
The team chooses the region, configuration and permissions, deploys the workload and monitors consumption, availability and security. The provider manages part of the physical infrastructure, but the customer remains responsible for aspects such as configuration, data and access control, depending on the service contracted.
Practical example
A small company creates an internal dashboard to track orders. At first, it can use an application hosted on a managed platform, with a database and file storage provided as services. If demand grows, the team can adjust capacity without buying new equipment.
This flexibility does not eliminate planning. The company must choose where data will be stored, restrict accounts and permissions, keep backups and test how to recover the service. It also needs to estimate usage and watch the provider reports, because contracted resources can stay active even when no one is using the application.
For an AI task, the same idea allows running a service that calls models or processes files without maintaining its own physical machine. The architecture still needs to protect data, limit access and define what to do when an external service becomes unavailable.
What it's for
Cloud computing is used to host websites, store files, run applications, analyze data, collaborate in online tools and provide temporary processing capacity. Organizations can start with few resources and adjust infrastructure as needs change, as long as they monitor performance and usage.
It is not always simpler, safer or cheaper by definition. A poor configuration can expose data, create vendor lock-in or keep unnecessary resources running. To decide, compare requirements for control, privacy, availability, data location, portability and the team's ability to operate the solution.
The core idea is to consume infrastructure as a service, instead of managing each physical component. The model transfers part of the operation to the provider, but it does not automatically transfer the responsibility to design, configure and monitor the system.