What is cloud computing? A plain-English explanation

The problem cloud computing solves

Before the cloud, running software meant owning infrastructure. A company that wanted a website or an internal application had to buy physical servers, find somewhere to put them, connect them to power and networking, and employ people to keep them running. All of that was capital expense — large sums spent up front, before the software earned anything.

Worse, capacity had to be guessed in advance. Buy too little hardware and the application falls over the first time demand spikes; buy too much and expensive machines sit idle for years. And provisioning was slow: getting a new server ordered, delivered, racked and configured could take weeks or months, which meant every new idea waited on hardware.

Cloud computing dissolves all three problems at once. Capital expense becomes operating expense — you pay a provider as you go rather than buying up front. Capacity guessing disappears because you can add or remove resources in minutes as demand changes. And provisioning collapses from weeks to moments, because a new server is now an API call, not a purchase order.

How it actually works

Behind the word “cloud” sits something very physical: enormous data centres full of servers, owned and operated by the provider, connected to the internet. The provider’s trick is sharing that hardware safely and efficiently among many customers, which rests on a technology called virtualisation — software that divides one physical machine into many isolated virtual machines, each behaving like a computer of its own.

On top of virtualisation sits self-service. You do not phone the provider to ask for a server; you request one through a web console or an API, and it exists moments later. Everything is programmable: creating machines, attaching storage, configuring networks and deleting it all again can be done by scripts as easily as by humans. Metering completes the picture — the provider measures exactly what you consume and bills you for that, so the economics track your actual usage rather than your worst-case forecast.

The defining characteristics

Not everything hosted on the internet is cloud computing. The term has a reasonably precise meaning, defined by a handful of characteristics that a genuine cloud service exhibits:

• On-demand self-service — you provision resources yourself, instantly, without human intervention from the provider.
• Elasticity — capacity scales up and down with demand, sometimes automatically, so you are never locked to a fixed size.
• Pay-as-you-go pricing — you pay for what you use, not for what you own, and stopping a resource stops its cost.
• Measured service — usage is metered transparently, so consumption and cost are visible and auditable.
• Broad network access — everything is reachable over standard networks from ordinary devices, anywhere with connectivity.

The major providers

Three companies dominate public cloud computing: Amazon Web Services (AWS), Microsoft Azure, and Google Cloud. Each operates global networks of data centres and offers broadly comparable building blocks — computing, storage, databases, networking — alongside hundreds of higher-level services, from machine learning to data analytics.

The concepts in this article apply to all of them, which is worth emphasising: cloud computing is an industry model, not a product of any single vendor. That said, each provider has its own service names, consoles and certification path, so most people pick one to learn deeply first. AWS is the most common starting point simply because of how widely it is used, and its foundational certification — the Cloud Practitioner — is built around exactly the concepts covered here.

The honest limitations

Cloud computing is genuinely transformative, but it is not automatically cheaper, simpler or better for every workload, and pretending otherwise sets beginners up badly. Pay-as-you-go pricing cuts both ways: without governance, forgotten resources and unexpected usage patterns produce surprising bills, and cost management is a discipline cloud teams must actively practise.

There is also dependency to weigh. Building deeply on one provider’s services makes leaving that provider harder — a trade-off usually worth making for the productivity gained, but a trade-off nonetheless. Everything depends on network connectivity, which matters for some environments more than others. And a steady, predictable, always-on workload can sometimes be cheaper on owned hardware than rented capacity. The cloud’s strength is flexibility; workloads that need none of it benefit least.

Why this is where certification study begins

Every cloud certification, on every provider, starts from the ideas in this article: what the cloud is, why organisations adopt it, the shift from capital to operating expense, elasticity, and pay-as-you-go economics. The AWS Certified Cloud Practitioner devotes an entire exam domain to cloud concepts, and the associate-level exams — Solutions Architect included — assume this foundation completely and build scenario questions on top of it.

If these ideas are new to you, they are the right place to start, and they reward being learned properly rather than memorised. From here, the natural next concepts are the service models — IaaS, PaaS and SaaS, which describe how much of the stack the provider manages — and the deployment models of public, private and hybrid cloud. Both are explained in our companion articles, and both sit alongside this material in the opening chapters of any cloud certification syllabus.