AWS Global Infrastructure
Regions, Availability Zones, and edge locations — how AWS is physically organized and why it matters for availability and latency.
The building blocks
AWS's physical footprint is organized in a strict hierarchy: Regions contain Availability Zones, which contain one or more data centers. Separately, a much larger network of edge locations brings content close to end users. Nearly every architecture and availability question on the exam builds on these terms.
| Component | What it is | Key numbers/facts |
|---|---|---|
| Region | A physical geographic area (e.g., us-east-1 in N. Virginia) containing multiple isolated AZs | 30+ worldwide; you choose where data lives |
| Availability Zone (AZ) | One or more data centers with independent power, cooling, and networking | Each Region has at least 3 AZs, linked by low-latency private fiber |
| Edge location | A small site that caches content near users, used by CloudFront and Route 53 | 400+ locations — far more than Regions |
| Local Zone | An extension of a Region placing compute/storage in a large city for single-digit-ms latency | For latency-critical apps near specific metros |
| AWS Wavelength | AWS compute embedded in telecom 5G networks | Ultra-low latency for mobile/5G apps |
Think of a Region as a metropolitan area, each AZ as a separate self-sufficient campus across town from the others, and edge locations as thousands of corner kiosks worldwide that keep popular items in stock so nobody drives across the country for them.
Choosing a Region
Four factors when picking a Region
- Compliance / data residency — legal requirements may dictate where data must physically stay. This factor overrides all others.
- Latency (proximity to users) — closer Regions mean faster responses.
- Service availability — new services and features don't launch in every Region at once.
- Pricing — the same service can cost different amounts in different Regions.
Deploying across multiple AZs = high availability within a Region. Deploying across multiple Regions = disaster recovery and lower global latency. Data does not leave a Region unless you configure it to (e.g., replication) — a fact AWS emphasizes for data residency.
The edge network
AWS's CDN: caches your content (images, video, APIs) at edge locations so users load it from nearby instead of from your Region.
Global DNS service that answers queries from the edge network, routing users to healthy, nearby endpoints (100% availability SLA).
Routes user traffic onto AWS's private backbone at the nearest edge for faster, more stable performance to your app — no caching involved.
CloudFront caches content; Global Accelerator accelerates the network path (static IPs, no caching). If the scenario is about media/static files, choose CloudFront; if it's about improving global performance of a TCP/UDP application, choose Global Accelerator.
What is an AWS Availability Zone?