What is Amazon EC2? A plain-English explainer
Amazon EC2 (Elastic Compute Cloud) is AWS’s core compute service: it gives you resizable virtual servers, called instances, that run in AWS’s data centres instead of your own. Rather than buying a physical machine, racking it and maintaining it, you launch an instance in minutes, use it for as long as you need, and pay for the capacity you actually consume. EC2 is infrastructure-as-a-service (IaaS) — AWS owns and operates the underlying hardware, and you get a virtual machine you control from the operating system upward. It is one of the oldest and most widely used AWS services, and because it embodies the core cloud-computing idea — capacity on demand, no upfront hardware — it is also one of the first services every cloud certification introduces. Here is what an instance actually is, what EC2 is used for, and how it relates to the newer compute options that sit alongside it.
What an “instance” actually is
An EC2 instance is a virtual server: a slice of a physical host machine, made to behave like an independent computer with its own operating system, storage and network connection. Under the bonnet, AWS uses virtualisation to carve one physical server into many isolated virtual ones — our what-is-virtualization explainer covers that mechanism in more depth, but the practical effect is what matters here: you get a machine that behaves like dedicated hardware without AWS having to give you an entire physical server to yourself.
You choose the operating system (various Linux distributions or Windows Server), and from that point it behaves like any server you have full control over — you can install software, open network ports, attach storage, and run whatever workload you like. What makes it “elastic” is that you are not locked into that choice of capacity: you can resize, replace or add more instances as demand changes, which is the whole difference from buying a fixed physical machine upfront.
That resizing is genuinely fast, too — launching, stopping or replacing an instance is a matter of minutes rather than the weeks a physical server order and installation would take, which is a large part of why EC2 changed how teams plan for capacity in the first place.
What EC2 is used for
Because an instance is a general-purpose virtual server, EC2 can run almost anything: web applications and APIs, background processing jobs, databases you choose to manage yourself, development and test environments, or entire legacy systems migrated from a physical data centre. It is the closest cloud equivalent to “just give me a server”, which is exactly why it remains the default starting point for workloads that do not fit neatly into a more specialised managed service.
A common pattern is hosting a website or application tier on one or more instances, often placed behind a load balancer so traffic spreads across them and a failed instance does not take the whole application down. Development and test teams also use EC2 to spin up disposable environments that mirror production, then shut them down when the work is done — something a physical server could never do on demand.
How you pay for it, without the numbers
EC2 follows the pay-for-use model that runs through most of AWS: you are billed for the capacity and time you consume, rather than paying a fixed amount regardless of use. Instance sizes and pricing options vary and change often, so this article deliberately does not list them — check AWS’s EC2 pricing page for the current detail. The concept to take away is that stopping or resizing instances you are not using is how EC2 stays cost-efficient; capacity left running idle is capacity you are still paying for.
There is also a limit to how much capacity an account can launch by default, as a safeguard against runaway spend — check AWS’s documentation if you expect to need more than the starting allowance.
How EC2 fits with serverless and containers
EC2 sits at one end of a spectrum of AWS compute options. With EC2, you manage the operating system, patching and scaling yourself — maximum control, more operational responsibility. AWS Lambda, at the other end, is serverless: you supply code and AWS runs it without you provisioning any server at all (see our what-is-serverless-computing explainer). Containers occupy the middle ground — a lighter-weight way to package and run applications that can run on EC2 instances themselves or on serverless container platforms; our containers-vs-virtual-machines explainer covers how that packaging differs from a full virtual machine like an EC2 instance.
None of these options replaces the others outright. Many real AWS architectures mix them: EC2 for workloads that need full control or specific software, Lambda for event-driven bursts of work, containers for portable application packaging. Choosing between them is a recurring theme across AWS’s architecture-focused certifications.
Where EC2 appears in certification study
EC2 is foundational enough that it appears in some form on every major AWS exam. AWS Cloud Practitioner introduces it as the canonical example of IaaS and elastic capacity. AWS Certified Solutions Architect – Associate goes considerably further: instance purchasing options, placement, scaling and how EC2 combines with storage, networking and load balancing to form resilient architectures. AWS Certified CloudOps Engineer – Associate focuses on the operational side — monitoring, patching and managing EC2 fleets once they are running.
This article covers what EC2 is and why it exists — the detail of instance families, purchasing options and architecture patterns belongs in the syllabus itself. Our /revision study library breaks that down lesson by lesson for each of these certifications.
Original practice questions, timed mock exams and revision notes. No card, nothing to pay.