What is Amazon S3? A plain-English explainer
Amazon S3 (Simple Storage Service) is AWS’s object storage service: it lets you store and retrieve any amount of data — files of essentially any type or size — as “objects” held inside containers called “buckets”. There is no server to provision and no capacity to plan for in advance; you create a bucket, put objects into it, and S3 scales to hold however much you give it. It is designed for extremely high durability, meaning AWS engineers it so your data is very unlikely to be lost, though the exact guarantee is best read on AWS’s own S3 page rather than repeated as a fixed figure here. S3 is one of the oldest AWS services and one of the most widely used, because “store this file somewhere reliable and reachable” is a need almost every application has. Here is what an object actually is, what people use S3 for, and how it fits with other AWS storage options.
What object storage actually is
S3 stores data as objects rather than as files on a disk or blocks on a volume. An object is simply the data itself — a photo, a video, a backup archive, a log file — plus some metadata, stored flat inside a bucket and addressed by a unique key rather than a folder path. That is a different model from the file systems most people are used to, and from the block storage that underpins a database’s disk; our object-vs-block-vs-file-storage explainer covers how the three compare and when each is the right fit.
Practically, this flat, key-based model is what lets S3 scale so far without you managing any of the underlying infrastructure. You never provision disk space or worry about running out of room on a volume — you simply keep writing objects, and AWS handles where and how they are physically stored.
Buckets, objects and keys
A bucket is a container for objects — think of it as a top-level namespace with a globally unique name. Inside a bucket, every object is identified by a key, which can look like a file path (for example, “images/2026/logo.png”) even though S3 is not a hierarchical file system underneath; that layout is a convenience for organising and browsing, not a real folder structure.
Buckets also carry configuration that applies to everything inside them: who can access the objects, whether versions of an object are kept as it changes, and which storage class new objects use by default. Getting that configuration right — particularly access permissions — is one of the most common real-world S3 tasks and a frequent exam topic, because a misconfigured bucket is a well-known source of accidental data exposure.
What people use S3 for
Because object storage is so general-purpose, S3 shows up in almost every kind of application:
- Backups and archives — a durable, off-site place to keep copies of data that must survive hardware failure.
- Media storage — images, video and audio for websites and applications, often paired with a content delivery network.
- Data lakes — the raw storage layer underneath big-data and analytics tools, holding structured and unstructured data together.
- Static website hosting — S3 can serve the plain HTML, CSS, JavaScript and images of a website directly, with no web server to run.
- Application assets — anything an application needs to read or write that is not suited to a database, from uploaded user files to generated reports.
Storage classes, at a high level
Not all data needs to be equally quick to retrieve. S3 offers several storage classes built around that trade-off: some are optimised for data you access constantly, others for data you rarely touch but must still keep, at a lower ongoing cost in exchange for slower or less frequent retrieval. This article deliberately does not list specific classes or their pricing, because the line-up and rates change over time — AWS’s S3 storage classes page is the accurate, current source. The concept worth remembering is that matching the storage class to how often you actually access the data is the main lever for keeping S3 costs sensible.
How S3 fits with the rest of AWS
S3 is pay-for-use like most of AWS: you pay for the storage you consume, plus the data you move in and out, rather than a fixed capacity you had to guess in advance. It also acts as connective tissue for a lot of the platform — analytics services read data lakes built on S3, machine-learning services train on data stored there, and content delivery networks cache S3-hosted content close to users. Very few substantial AWS architectures have no S3 in them at all.
Where S3 appears in certification study
S3 is core material on AWS Cloud Practitioner, which introduces object storage and buckets as foundational AWS concepts. It goes deeper on AWS Certified Solutions Architect – Associate, covering storage classes, access control and how S3 combines with other services in real architectures, and on AWS Certified Developer – Associate, which focuses on interacting with S3 programmatically from an application. Our /revision study library covers that depth lesson by lesson.
Original practice questions, timed mock exams and revision notes. No card, nothing to pay.