AWS
Also known as: Amazon Web Services, Amazon Cloud
AWS is Amazon's cloud computing platform.
Instead of buying and maintaining your own servers, you rent computing power, storage, and services from Amazon's global data centres: paying only for what you use.
AWS, explained at three levels
At home
Imagine needing a server to run your website, but instead of buying a physical computer, you rent one from Amazon and pay by the hour. That's AWS: a massive collection of computing resources you can access over the internet without owning any hardware.
At work
AWS is a catalogue of rented computing services: virtual servers, storage, databases, and analytics that a business switches on through a website or code and pays for by the hour or gigabyte. A company avoids buying hardware up front, can add capacity in minutes when demand jumps, and shuts it off when demand drops, turning a large capital purchase into a monthly operating bill.
At home
AWS is built from separate services you combine like building blocks. EC2 gives you virtual machines, S3 stores files as objects, RDS runs managed databases such as PostgreSQL, and Lambda runs code only when an event fires. You control them through the web console, a command-line tool, or an API, and every action is an authenticated API call underneath. AWS handles the hardware, power, cooling, and replacing failed disks.
At work
An AWS account is organized around regions, each containing several availability zones, and around a few control layers. IAM defines who or what can call each API action. A VPC is your private network slice, split into subnets, with security groups acting as per-server firewalls. Auto Scaling and Elastic Load Balancing add or remove servers as traffic changes, and billing metres each resource separately, so architecture choices directly shape the monthly cost.
At home
Under the hood, AWS is a set of regions, each made of availability zones: clusters of data centres with separate power and cooling, linked by fast private fibre. Most services are regional, so surviving a zone failure is something you design for by spreading servers across zones. Higher-level services like Lambda and Fargate hide the virtual machine entirely, trading control for convenience. In my view, that trade is worth it for small projects, but the bill becomes harder to predict.
At work
Enterprise AWS is mostly about governance. AWS Organizations groups many accounts under one payer, and service control policies cap what any account can do, which limits the damage of a mistake. CloudTrail logs every API call for audit, CloudWatch collects metrics and alarms, and CloudFormation, CDK, or Terraform define infrastructure as reviewable code. The main failure modes are overly broad IAM permissions, single-region designs, and runaway costs, which Savings Plans and budgets only partly control.
Real-world examples
At home
Sophie uses AWS S3 to host her photography portfolio as a static website. She pays less than a dollar a month because S3 charges per gigabyte stored and per request, and her traffic is light.
At work
Thomas's team deploys their SaaS application on AWS using EC2 auto-scaling groups behind an Application Load Balancer, with RDS PostgreSQL for the database and CloudFront as a CDN. The whole infrastructure is defined in Terraform so any developer can spin up a staging environment in minutes.
Common confusions
- AWS is just for big companies — The pay-as-you-go model and free tier make AWS accessible to individual developers and startups. Many solo projects run for pennies a month.
- AWS and Amazon.com are the same thing — AWS is a separate business unit. Amazon.com (the store) runs on AWS, but AWS also powers Netflix, Airbnb, NASA, and thousands of other organizations.
- Moving to AWS automatically makes your system reliable — AWS provides the building blocks for reliability, but you must architect for it: using multiple availability zones, health checks, and proper backup strategies.
For analysts
Business analysts use AWS QuickSight for dashboards, Glue for data pipelines, and Athena to query S3 data with SQL: no server provisioning required.
History & origin
Amazon launched AWS's core services in 2006 with just a handful of services: S3 for storage and EC2 for virtual servers. It was originally built to support Amazon's own e-commerce infrastructure. When engineers realized other companies would pay for the same capacity, AWS became a separate business. Today it holds roughly 30 percent of the global cloud infrastructure market and usually generates most of Amazon's operating profit.