Cloud regions, availability zones and edge locations
In short: Providers group data centres into availability zones and zones into regions. Spreading a workload across zones protects it from a single site failure, choosing the right region helps with latency and data residency, and edge locations bring content closer to users.
How the global footprint is organised
A cloud provider does not run one giant data centre. It runs many, grouped so that customers can design around failures. The names differ slightly between providers, but the ideas are the same.
- Region: a geographic area, such as a city or metro area, with its own independent set of services. You pick a region when you create most resources.
- Availability zone (AZ): one or more data centres inside a region with separate power, cooling and networking. Zones are far enough apart that a fire or flood should not hit two at once, but close enough for low-latency links.
- Edge location: a smaller site used by content delivery networks (CDNs), DNS and some security services to serve users from nearby.
Not every region has the same services or the same number of zones, so always check before you design.
Designing for resilience
If every server sits in one zone, a single power or network failure takes your service down. A multi-AZ design runs copies in at least two zones behind a load balancer, with databases replicating between them. This handles most hardware and site failures. A multi-region design goes further and survives the loss of a whole region, but it is more complex and more expensive because data must be copied across long distances. See high availability and disaster recovery for the patterns.
Latency, data residency and sovereignty
| Concern | What to consider |
|---|---|
| Latency | Choose a region near most users; use a CDN for static content |
| Data residency | Laws or contracts may require personal data to stay in a country or bloc |
| Sovereignty | Some customers also want local operation and control, not just local storage |
| Cost | Prices differ by region, and moving data between regions is charged |
| Service availability | New services often launch in a few regions first |
Data residency is about where data is stored. Remember that backups, logs and replicas also count. A database in a Middle East region whose backups copy to Europe may break a residency rule.
Security practices
- Agree an approved list of regions with your compliance team.
- Block other regions with a policy (for example an AWS service control policy or an Azure Policy "allowed locations" rule).
- Turn on logging in every region you allow, not just your main one.
- Check where backups and replicas go.
- Test a zone failure in a lab to prove failover works.
Tip: attackers who steal cloud keys sometimes start resources in regions you never use, hoping nobody looks there. Blocking unused regions closes that gap.
Practice
What is the difference between a region and an availability zone?
Your web app runs on two servers in the same zone. What is the risk?
Why should backups be included in a data residency review?
Key takeaways
- A region holds several isolated availability zones.
- Multi-AZ designs survive the loss of one data centre site.
- Edge locations serve content and DNS close to users.
- Pick regions for latency, residency, cost and service availability.
- Block unused regions and log every region you allow.
Original content written by Mohamed Fayas for FM TecHub. © 2026 FM TecHub. You're welcome to share a link to this page; please don't copy the text elsewhere without permission.