What is a data centre?
A data centre (also spelled "data center" — this article uses the British spelling, though the American spelling is the more common search term) is a purpose-built facility that houses the physical servers, storage systems, networking equipment and the power and cooling infrastructure needed to keep them all running. Every website, app and cloud service ultimately runs on hardware sitting in one of these buildings somewhere — the cloud does not remove the data centre, it just means you rent capacity inside someone else’s instead of building your own. This guide explains what is actually inside one, how cloud providers organise data centres into regions and availability zones, and why "the cloud" is really just a very large, very automated way of renting data-centre capacity.
What is physically inside a data centre
At its core, a data centre is racks of servers — the physical computers that store data and run applications — connected by high-speed networking, sitting on raised floors or in purpose-built rooms designed around three problems: power, cooling and physical security. Servers generate significant heat and need a constant, reliable electricity supply, so a serious data centre includes backup generators and battery systems, and dedicated cooling systems to keep the hardware within safe operating temperatures. Physical security — controlled access, monitoring, and restrictions on who can even walk up to the hardware — matters because someone with physical access to a server can bypass a great deal of software-level protection.
Data centre tiers, as a concept
Data centres are sometimes described using a "tier" concept that reflects how redundant and fault-tolerant the facility is — broadly, higher tiers add more redundant power paths, cooling systems and backup capacity, so the facility can keep running through more kinds of equipment failure or maintenance without an outage. The exact tier definitions and numeric uptime targets vary by rating body and change over time, so this article deliberately does not restate specific numbers — the concept to take away is that redundancy, not raw size, is what separates a basic facility from a highly resilient one.
How cloud regions and availability zones map to data centres
Cloud providers do not run one giant data centre per region — they group multiple physically separate data centres into what they call an availability zone, and group multiple availability zones, spread across a geographic area, into what they call a region. The idea is straightforward: by spreading availability zones across separate facilities with independent power and networking, a fire, power outage or other failure at one data centre should not take down an entire region. Our explainer on AWS regions and availability zones covers this structure in more depth — the same layered idea (data centres, grouped into zones, grouped into regions) applies, with different naming, across the other major cloud providers too.
Why the cloud abstracts the data centre away
Before cloud computing, running an application meant either building your own data centre or renting dedicated space and hardware in someone else’s (co-location) — a large upfront commitment regardless of how much capacity you actually needed. Cloud computing rents you compute, storage and networking as a service, so the underlying data centre — which rack, which building, which cooling system — becomes something the provider manages and you never need to think about. Our what is cloud computing explainer covers that shift from owning infrastructure to renting it in full; the data centre itself does not disappear, it just moves behind an interface you access over the internet instead of a building you have to visit.
Data centres versus "the cloud" — they are not opposites
It is easy to talk about "on-premises" and "the cloud" as if the cloud has no physical location — it does. Cloud computing is data centres, at a scale and level of automation most individual organisations could never justify building themselves, made available on demand. The meaningful contrast is not "data centre versus cloud," it is "a data centre you own and operate yourself versus capacity you rent inside a provider’s data centres" — the same underlying hardware, a very different operating model.
Where this appears in cert study
Foundational certifications such as AWS Cloud Practitioner and Microsoft Azure Fundamentals introduce regions, availability zones and the shared-responsibility model that follows from the provider owning the physical data centre while the customer is responsible for what they run inside it. Google Cloud’s Associate Cloud Engineer exam covers the equivalent regional/zonal structure on Google Cloud. None of these exams expect you to know a specific data centre’s exact layout — the concept of physical resilience through separation is what gets tested.
Original practice questions, timed mock exams and revision notes. No card, nothing to pay.