AWS global infrastructure — Regions, Availability Zones and edge locations
AWS runs its services from data centers all over the world. They are organized in layers: Regions contain Availability Zones, and Availability Zones contain data centers. Separately, edge locations bring some services physically close to users.
Understanding these layers is the key to two big ideas on the exam: high availability and choosing where your data lives.
Regions
A Region is a physical area of the world where AWS has several Availability Zones, such as Frankfurt (eu-central-1), Tokyo (ap-northeast-1) or Northern Virginia (us-east-1).
- Regions are independent of each other. A problem in one Region does not spread to another.
- Your data stays in the Region you choose unless you copy or move it yourself. This matters for laws about where data may be stored.
- Most services are regional: an EC2 instance or an S3 bucket lives in one Region. A few are global, such as IAM and CloudFront.
Availability Zones
An Availability Zone (AZ) is one or more discrete data centers with their own redundant power, networking and connectivity.
- Every Region has at least three AZs.
- AZs are physically separated by many kilometers, but all within 100 km (60 miles) of each other, so they are close enough for very fast connections.
- AZs do not share single points of failure: a fire, flood or power cut in one should not affect the others.
Running copies of your application in two or more AZs is the standard way to make it highly available. If one AZ has trouble, the copies in the other AZs keep serving users.
Edge locations
Edge locations are sites in many more cities than there are Regions. They do not run your servers; instead, services use them to be close to users:
- Amazon CloudFront caches copies of your content (images, videos, web pages) so users download them from nearby instead of from a distant Region.
- Amazon Route 53 answers DNS queries from edge locations, so names resolve quickly.
- AWS Global Accelerator uses the edge network to route users onto AWS’s fast private network sooner.
Two related options bring AWS closer still: AWS Local Zones place selected services near large cities, and AWS Outposts puts AWS hardware in your own data center.
When to use more than one Availability Zone or Region
| Goal | Use | Why |
|---|---|---|
| Survive a data center or AZ failure | Multiple AZs in one Region | AZs fail independently; this is the usual high-availability setup |
| Disaster recovery and business continuity | Multiple Regions | A copy far away survives even a regional problem |
| Low latency for users on different continents | Multiple Regions (and CloudFront) | Each user is served from somewhere close |
| Data sovereignty | A specific Region | Laws or contracts may require data to stay in one country |
How to choose a Region
Ask four questions, in roughly this order:
- Compliance and data residency. Does a law or contract require the data to stay in a certain country? If so, that decides it.
- Proximity to users. Which Region gives the lowest latency for most of your users?
- Service availability. Does the Region offer every service and feature you need? New services do not launch everywhere at once.
- Pricing. The same service can cost different amounts in different Regions.
Putting it together: a European online shop
A shop based in Germany sells across Europe. Customer data must stay in the EU.
- It chooses Frankfurt (
eu-central-1) for compliance and closeness to customers. - It runs its web servers and database in three Availability Zones, so losing one data center does not stop sales.
- It uses CloudFront, so product images load quickly from edge locations in many European cities.
- For disaster recovery it keeps backups in a second EU Region, such as Ireland.
Common mix-ups
- An AZ is not one data center. It is one or more data centers.
- An edge location is not a Region. You cannot launch an EC2 instance in an edge location; it serves cached content and DNS.
- Multi-AZ vs multi-Region. Multi-AZ is for high availability within a Region. Multi-Region is for disaster recovery, global users or data rules.
Check yourself
A company must keep all customer data inside Canada. What is the first thing that decides which Region it uses?
Compliance and data residency. It must pick a Region in Canada before thinking about latency, services or price.
How do you make an application keep running if one data center loses power?
Run it in more than one Availability Zone. AZs have separate power and networking, so the copies in other AZs keep working.
Users in Australia complain that videos hosted in a US Region load slowly. What helps without moving the whole application?
Amazon CloudFront. It caches the videos at edge locations near the users.
Key takeaways
- Regions contain at least three Availability Zones; AZs contain one or more data centers.
- Multiple AZs give high availability; multiple Regions give disaster recovery, global reach and data residency options.
- Edge locations serve cached content and DNS close to users through services like CloudFront and Route 53.
Cloud Study Lab is an independent study site. It is not affiliated with or endorsed by Amazon Web Services. AWS and the names of AWS certifications are trademarks of Amazon.com, Inc. or its affiliates.