AWS Global Infrastructure

Regions, Availability Zones, edge locations, and how to pick where your workload runs.

What is it?

AWS runs its services from physical locations spread around the world. Knowing the vocabulary is the first step to building something that is fast, legal, and survives failures.

  • Region: a separate geographic area (for example us-east-1 or eu-west-1) made up of multiple data center clusters. Regions are isolated from one another; your data stays in a Region unless you move it.
  • Availability Zone (AZ): one or more discrete data centers inside a Region, with independent power, cooling, and networking, connected to the other AZs by fast, low-latency links. Region names plus a letter identify them, like us-east-1a.
  • Edge location: a site used by services such as CloudFront (CDN) and Route 53 to deliver content closer to end users. There are many more edge locations than Regions.
  • Local Zones: small extensions of a Region placed near large cities for single-digit-millisecond latency.
  • Outposts: AWS-managed hardware installed in your data center, giving you AWS APIs on-premises.

Choosing a Region comes down to four factors, usually considered in this order:

  • Compliance: laws or policies about where data may live come first and are not negotiable.
  • Latency: choose a Region near most of your users.
  • Service availability: not every service or feature launches in every Region at once.
  • Pricing: the same service costs different amounts in different Regions.

More building blocks of the global network

  • AWS Wavelength: embeds AWS compute and storage inside 5G telecom networks, so mobile apps (game streaming, live video, connected vehicles) reach servers with very low latency.
  • AWS Local Zones: Region extensions placed near big cities for single-digit-millisecond access; you opt in and use them like a subnet of the parent Region.
  • AWS Global Accelerator: gives your app two static anycast IP addresses and routes users over the AWS backbone to the nearest healthy endpoint. Unlike CloudFront it does not cache content; it speeds up and stabilises TCP/UDP traffic and fails over automatically.

Availability Zones are one or more discrete data centers with separate power, networking and connectivity, linked by fast low-latency links. They are placed far enough apart that a single event is unlikely to hit two of them, so they share no single point of failure. Running across at least two AZs is the usual high-availability design.

Why use more than one Region? (1) Disaster recovery and business continuity if a whole Region is impaired, (2) lower latency for users on other continents, (3) data sovereignty and compliance rules that require data in a given country or area, and (4) access to services that launch in some Regions first. Data never leaves a Region unless you choose to copy it.

Explain like I'm 10

Think of a worldwide delivery company. A Region is a whole city's logistics network. An Availability Zone is one warehouse district within that city, with its own power and roads, so a blackout in one district does not stop the others. Edge locations are neighborhood pickup lockers that hold popular parcels close to customers.

Examples

Exploring Regions and AZs with the CLI

# All Regions enabled for your account
aws ec2 describe-regions --query "Regions[].RegionName" --output text

# AZs in one Region
aws ec2 describe-availability-zones \
  --region eu-west-1 \
  --query "AvailabilityZones[].[ZoneName,State]" \
  --output table

# Make a Region your default for this profile
aws configure set region eu-west-1

Most CLI calls are Region-scoped. Forgetting which Region you are in is a classic source of 'where did my server go?'

Region-choice worksheet

Workload: customer portal for German retail customers

1. Compliance : customer data must stay in the EU        -> EU Regions only
2. Latency    : users are in Germany                     -> eu-central-1 (Frankfurt)
3. Services   : needs a service launched there?          -> verify in the docs
4. Price      : compare with eu-west-1 if latency is OK  -> decide

How it works

Each Region contains at least three AZs in most cases, placed far enough apart to avoid sharing a single disaster (flood, fire) but close enough for synchronous replication. When you launch a resource you pick a Region, and for many resources an AZ (or a subnet that lives in one AZ).

Designing for high availability means spreading copies of your workload across multiple AZs. Designing for disaster recovery or global users may mean using multiple Regions. Edge locations are separate: they cache and accelerate rather than run your servers.

Some services are global (IAM, Route 53, CloudFront configuration) while most are Regional. Data does not leave a Region unless you configure replication or copy it.

                AWS Global Infrastructure
 +------------------------ Region (eu-west-1) -------------------+
 |  +-- AZ a --+   +-- AZ b --+   +-- AZ c --+                    |
 |  |  data    |   |  data    |   |  data    |   <- low-latency   |
 |  |  centers |===|  centers |===|  centers |      private links |
 |  +----------+   +----------+   +----------+                    |
 +--------------------------------------------------------------- +
        ^                                   ^
        |                                   |
  [Local Zone]                       [Edge locations]
  near a big city                    cache content near users
        |
  [Outposts] = AWS hardware in your own data center

Why does it exist?

A single data center is a single point of failure, and a single country is a long way from users elsewhere. Splitting capacity into isolated Regions and AZs lets you survive failures, meet data-residency laws, and keep latency low.

When to use it

Use multiple AZs for almost any production workload that must stay up. Use multiple Regions when you need disaster recovery from a Region-wide event, very low latency for users on different continents, or a legal requirement for separate geographies. Use Local Zones or Outposts when milliseconds matter or data must stay on your premises.

When not to use it

Do not go multi-Region by default - it adds cost, data-replication complexity, and operational burden. For most apps, multi-AZ in one Region is the right first step. Do not pick a Region just because it is the console default.

Common mistakes

  • Choosing a Region only by price, ignoring compliance and latency.

  • Putting everything in one AZ and calling it 'highly available'.

  • Confusing an edge location with an AZ - edge locations do not host your EC2 instances.

  • Looking for a resource in the console and not seeing it because the Region selector is set differently.

  • Assuming AZ names (like us-east-1a) map to the same physical data center for every account; they are mapped per account.

  • Thinking Global Accelerator and CloudFront are the same: CloudFront caches content at the edge, Global Accelerator accelerates traffic to your own endpoints.

  • Believing that two resources in the same AZ are highly available - spread them across AZs.

Practice exercises

  1. Easy:

    In your own words, explain the difference between a Region, an Availability Zone, and an edge location.

  2. Easy:

    List the four factors for choosing a Region and rank them for an app that stores health records for Canadian patients.

  3. Medium:

    Run the describe-availability-zones command above for two Regions. How many AZs does each have?

  4. Hard:

    Design (on paper) a deployment that survives the loss of one AZ and then extend it to survive the loss of a whole Region. List what extra cost and complexity each step adds.

Interview questions

What is an Availability Zone?

One or more discrete data centers in a Region with independent power, cooling, and networking, linked to other AZs in the Region by low-latency connections.

How do you choose a Region?

Compliance and data residency first, then latency to users, then which services are available there, then price.

What is the difference between an edge location and a Region?

Regions host your workloads; edge locations are used by services like CloudFront and Route 53 to cache and deliver content close to end users.

Exam-style: A company needs to keep data in a specific country by law. What should it consider first? (A) Price (B) Compliance and data residency (C) Latency (D) Number of edge locations

B. Legal requirements outrank all other Region-selection factors.

Exam-style: Which AWS offering places AWS-managed infrastructure inside a customer's own data center?

AWS Outposts.

Exam-style: Which service delivers AWS compute inside 5G networks for ultra-low latency mobile apps?

AWS Wavelength.

Exam-style: Which service provides static anycast IPs and routes users to the nearest healthy endpoint over the AWS backbone?

AWS Global Accelerator.

Give three reasons to deploy in multiple Regions.

Disaster recovery, lower latency for distant users, and data-residency or compliance requirements.