Cloud: What It Is, What It Means, What It’s For, Definition and Examples

The cloud is a set of servers, software and data hosted in remote data centres and accessed over the Internet rather than from your own computer. When you open your email in Gmail, save a photo to Google Drive or stream a series, you are using the cloud.

For a company, it means renting computing power, storage space and applications, paying for them according to usage, without buying and maintaining its own infrastructure.

Cloud: meaning and definition

Cloud is borrowed from the everyday English word. In computing, the term refers to computing resources that sit not on the user’s device but in data centres run by a provider and reached over a network: servers, storage, databases, applications.

The reference definition is the one from NIST, the US standards institute, published in September 2011 in document SP 800-145. For NIST, cloud computing is a model that enables ubiquitous, on-demand network access to a shared pool of configurable resources. These resources can be provisioned and released rapidly, with minimal provider intervention.

The same document sets out five essential characteristics:

1. On-demand self-service: users provision servers or storage themselves, without going through an operator.
2. Broad network access: resources can be reached via the Internet from computers, smartphones and tablets.
3. Shared resource pool: the provider’s hardware serves multiple customers simultaneously and is allocated dynamically.
4. Rapid elasticity: resources scale up or down according to load, including automatically.
5. Measured service: consumption is monitored and billed according to actual use.

A physical server rented for a fixed fee, with no ability to scale, does not meet these criteria. In that case it is more accurate to speak of traditional hosting.

What is cloud computing?

‘Cloud’ and ‘cloud computing’ are often used interchangeably, but they refer to two slightly different things. The cloud is the set of remote resources. Cloud computing is the model through which these resources are delivered as a service, via the Internet and on a pay-as-you-go basis. The meaning of cloud computing is tied to computing services such as servers, storage, databases, networking and software.

In practice, programs and data no longer reside on a computer’s hard drive or on a company server: they run on a provider’s servers. That is why you can open the same file from your laptop at the office and from your smartphone on the train, and always find the latest version.

For a business, the change is first and foremost economic. IT spending shifts from upfront investment in hardware (CapEx) to variable operating cost (OpEx). The time needed to set up a new environment drops from weeks to a few minutes.

How does cloud computing work?

The cloud rests on three elements: distributed data centres, virtualisation and the network.

Data centres. Providers run buildings with thousands of servers, cooling systems, backup power and high-capacity connections. The large providers spread them across several regions of the world. This brings services closer to users and replicates data in different locations.

Virtualisation. A piece of software called a hypervisor divides each physical server into multiple virtual machines (VMs). Each VM behaves like a standalone computer, with its own operating system and applications, and is isolated from the others even though it shares the same hardware. It is this mechanism that allows the provider to serve many customers with the same infrastructure and to allocate resources in a matter of seconds.

Network and orchestration. Users reach services over the Internet, via browser, app or API. Orchestration software distributes the load across servers, spins up new resources when demand grows and moves workloads if a machine fails. This is why a well-designed cloud service stays available even when a single server stops working.

What the cloud is used for: examples

The cloud lets you use software, store data and run applications without owning the infrastructure that hosts them. We all use it, often without realising.

In everyday life

• Email such as Gmail or Outlook.com.
• Storage and synchronisation of files and photos with Google Drive, Dropbox, iCloud or OneDrive.
• Film and music streaming on platforms such as Netflix and Spotify.
• Online office suites such as Microsoft 365 and Google Workspace.

In business

• Cloud storage and backup: off-site copies of data, recoverable after a failure, theft or ransomware attack.
• Cloud servers: virtual machines hosting websites, e-commerce and management systems, with resources that scale up during peak periods.
• Subscription software: CRM, ERP, marketing automation, e-invoicing.
• Development and testing: environments created and decommissioned in minutes, without buying hardware.
• Data analytics and artificial intelligence: processing large volumes of data and training models that require computing power that is hard to sustain in-house.
• Remote working: access to company applications and documents from anywhere.

What are cloud services?

Cloud services are computing resources that a provider makes available over the Internet, usually on a subscription or pay-as-you-go basis: infrastructure, development platforms or ready-to-use applications. What sets them apart is how much of the system the provider manages and how much remains in the customer’s hands. This division gives rise to the service models.

What are the main cloud computing service models?

NIST defines three: IaaS, PaaS and SaaS. In recent years FaaS, also known as serverless, has been added. An analogy helps: it is like choosing between renting a plot of land to build on, a house to furnish or a house that is ready to move into.

IaaS – Infrastructure as a Service

The provider supplies the basic infrastructure: virtual servers, storage and networking. The customer installs and manages the operating system, middleware and applications. It is the most flexible model, but it requires more technical expertise. Examples: Amazon EC2, Google Compute Engine, Azure Virtual Machines.

PaaS – Platform as a Service

In addition to the infrastructure, the provider manages the operating system, runtime and development tools. Developers focus on code, without dealing with servers and updates. Examples: Heroku, Google App Engine, Azure App Service.

SaaS – Software as a Service

The user accesses a complete application through a browser or app, with nothing to install or maintain. The provider handles everything, from infrastructure to updates. It is the most widespread model: Salesforce, Slack, Microsoft 365 and Gmail are all SaaS services.

FaaS – Function as a Service (serverless)

The code is split into functions that run only when an event triggers them, such as a form submission or a file upload. Servers exist, but they are managed entirely by the provider, which scales them automatically. You pay only for execution time. Examples: AWS Lambda, Google Cloud Functions, Cloudflare Workers.

Who manages what in the different models

What is cloud infrastructure?

Cloud infrastructure is the set of physical and software components that makes it possible to deliver cloud services. It comprises four layers:

• Compute: processors and GPUs, made available as virtual machines or containers.
• Storage: block storage for VM disks, shared file storage and object storage for large volumes of unstructured data.
• Network: virtual private networks, load balancers, firewalls and connections between data centres.
• Virtualisation and management: hypervisors, orchestration tools and dashboards or APIs through which the customer configures resources.

All service models rest on this foundation. In IaaS the customer uses it directly; in PaaS and SaaS it remains hidden behind the platform or the application.

What types of cloud are there?

Four deployment models are distinguished according to who owns the infrastructure and who can use it.

Public cloud

The resources belong to an external provider and are shared among many customers. Each one works in an environment logically isolated from the others, following the principle of multi-tenancy. The main providers, known as hyperscalers, are Amazon Web Services, Microsoft Azure and Google Cloud. The public cloud requires no upfront investment and offers almost unlimited scalability.

Private cloud

The infrastructure is dedicated to a single organisation. It can sit in the company’s own data centre or with a provider that reserves it for the client. It offers greater control over data, configurations and compliance, at the cost of higher expense and expertise. It is often chosen by sectors with strict regulatory requirements, such as finance and healthcare.

Hybrid cloud

A hybrid cloud combines public cloud and private cloud, or on-premise infrastructure, connected so that data and applications can move from one environment to the other. A typical example: sensitive data stays in the private cloud, while the e-commerce site draws on additional public resources during traffic peaks, such as Black Friday.

Multicloud

Multicloud involves using services from several public providers, for example storage with one provider and data analytics with another. It reduces dependence on a single provider and allows the best service to be chosen for each need, but it increases management complexity. It can coexist with a hybrid architecture.

NIST also describes the community cloud: an infrastructure shared by organisations with common needs, such as public bodies or research institutes.

Advantages and limitations of the cloud

The cloud pays off when flexibility and speed are needed, but it is not the right choice for every workload.

Advantages

• Usage-based costs: no upfront hardware purchase; you pay for what you consume.
• Scalability and elasticity: resources follow demand, with no need to oversize infrastructure for peaks.
• Speed: new servers, test environments or applications are available within minutes.
• Business continuity: backups and replicas across several data centres simplify disaster recovery.
• Access and collaboration: data and applications can be reached from any location and device.
• Outsourced maintenance: updates to the hardware and to many software components are the provider’s responsibility.

Limitations

• Dependence on connectivity: without Internet access, or if a provider’s service goes down, access is interrupted.
• Latency: a distant data centre slows down applications that are sensitive to response times. Edge computing reduces the problem by processing data closer to where it is generated.
• Vendor lock-in: proprietary services and formats make switching provider expensive.
• Hard-to-predict costs: forgotten resources left running or oversized inflate the bill. The FinOps discipline exists precisely to keep them under control.
• Less direct control: the physical infrastructure is not in the company’s hands.

Is the cloud secure?

The cloud can be highly secure, but security does not depend on the provider alone. The shared responsibility model applies: the provider protects data centres, hardware, network and virtualisation, while the customer is responsible for data, accounts, permissions and configurations. Many incidents originate precisely on the customer side, from storage left public, weak passwords or access never revoked.

There are four basic measures:

• Encryption of data at rest and in transit.
• Identity and access management: multi-factor authentication and permissions limited to what each user needs.
• Backup and monitoring to detect anomalies and restore data.
• Configurations checked regularly, especially on exposed storage and databases.

Then there is the regulatory issue. If you process personal data of individuals in the European Union, you remain the data controller under the GDPR even when the data sits on the provider’s servers. The provider acts as data processor, and a contract compliant with Article 28 is required. You should also check where the data is stored: many providers let you choose regions within the EU.

Cloud, client-server model and on-premise: the differences

In the traditional client-server model of the Internet, the client sends a request and the server responds, for example by returning a web page. The cloud uses the same network architecture, but the server does more than respond. It runs applications, stores data and processes information on the user’s behalf, using virtualised, shared and elastic resources.

For a company, however, the most useful comparison is between cloud and on-premise, meaning owned infrastructure installed at its own premises.

Why is it called the ‘cloud’?

In early technical diagrams of networks and telecommunications, the external part of the network, which did not need to be detailed, was drawn as a cloud. As the Internet spread, the symbol came to denote the set of remote resources whose physical location the user neither sees nor needs to know.

And containers? Are they IaaS, PaaS, SaaS or FaaS?

A container packages an application together with its libraries and dependencies, so that it runs the same way on any machine. Unlike a virtual machine, it does not emulate an entire computer: it shares the kernel of the host operating system. This makes it lighter, and it starts in a few seconds. Docker is the most widely used tool for creating them, Kubernetes the most widely used for orchestrating them.

Containers are not a service model but a packaging technology used across several models:

• on IaaS virtual machines, managing everything in-house;
• on managed Kubernetes services, such as Amazon EKS, Google Kubernetes Engine and Azure Kubernetes Service, often called CaaS (Container as a Service);
• on PaaS or serverless platforms that run containers without exposing the servers, such as Google Cloud Run and Azure Container Apps.

Containers are also the foundation of cloud-native applications, designed as sets of small independent services that are updated and scaled separately.