Three geographic layers
A Region is a geographic area containing multiple Availability Zones, such as ap-southeast-1 (Singapore). An Availability Zone (AZ) is one or more isolated data centers with independent power and networking, connected to the other AZs in the Region through low-latency links. An Edge Location places CloudFront content and Route 53 DNS near users; it is not where you launch EC2.
flowchart TB R["Region ap-southeast-1<br/>Singapore"] --> A["AZ 1a<br/>data centers"] R --> B["AZ 1b<br/>data centers"] R --> C["AZ 1c<br/>data centers"] A --> EC2["EC2 + EBS<br/>live in one AZ"] E["Edge Location<br/>CloudFront cache"] --> U["Nearby users"] %% step R: You choose a Region first: it decides latency, price, available services and where data is stored. %% step A: Each Region has at least three AZs, physically separated. %% step B: AZs have separate power and networking, so one failing rarely affects the others. %% step C: They are close enough for fast synchronous replication between them. %% step EC2: An instance, its EBS volumes and its subnet sit in exactly one AZ. %% step E: Edge Locations cache content close to users; you do not run servers there. %% step U: Users get lower latency for cached static content and DNS answers.
Choosing a Region
Pick a Region by:
- Compliance and data residency — some data must stay in a country.
- Latency — closer to users means faster responses.
- Service availability — new services and instance types do not launch in every Region on day one.
- Price — the same instance type can cost different amounts in different Regions.
Some newer Regions are opt-in and must be enabled for the account before you can use them.
Resource scope
| Scope | Examples |
|---|---|
| Global | IAM users, roles and policies; Route 53 hosted zones; CloudFront distributions |
| Regional | VPC, S3 bucket (the name is globally unique, the data lives in one Region), Lambda functions, AMIs |
| Zonal (one AZ) | EC2 instance, EBS volume, subnet |
High availability places workloads in multiple AZs — two instances in two subnets in different AZs behind a load balancer — rather than trying to stretch one subnet or EBS volume across AZs. Disaster recovery for a whole-Region outage goes one step further and copies data or infrastructure to a second Region.
AZ names such as ap-southeast-1a are mapped per account; the letter a in two accounts need not indicate the same physical location. To compare across accounts, use the AZ ID (for example apse1-az1), which is the same everywhere.
Other locations
- Local Zones extend a Region into a metro area for single-digit-millisecond latency.
- Wavelength Zones place compute inside 5G carrier networks.
- Outposts put AWS hardware in your own data center.
You will not use these in the labs, but knowing they exist helps when a question mentions "very low latency to one city" or "AWS services on-premises".
| Command | Purpose |
|---|---|
aws ec2 describe-regions --output table |
List enabled Regions |
aws ec2 describe-availability-zones --region ap-southeast-1 |
List AZs (names and AZ IDs) in a Region |
aws configure set region ap-southeast-1 |
Set the default Region for the CLI |
Traps: launching EC2 "in an Edge Location"; assuming an EBS volume can attach to an instance in another AZ; assuming every service exists in every Region.