1.7 Regions, availability zones and edge locations
Describes the cloud landscape as of August 2026
What this is and why it exists
"The cloud" is buildings with addresses, and choosing which buildings run your workload is a real decision with three inputs: how far your users are, what the law says about where data lives, and what each location costs. This lesson gives you the geography.
The vocabulary
- Region — a geographic cluster of data centres sold as one location, like a provider's Mumbai region.
- Availability zone (AZ) — one or more data centres inside a region with independent power, cooling and networking.
- Edge location — a small site near users that serves cached content and absorbs traffic; not a place to run your application.
- CDN (content delivery network) — the network of edge locations serving your static content from near the reader.
- Data residency — where your data is physically stored.
- Latency — the round-trip time between your user and the machine that answers.
The mental model
Concentric circles of independence. Zones within a region are engineered to fail independently — separate power, separate cooling — so an application spread across two zones survives a zone's failure. The honesty note: they share the region's control systems, and region-wide incidents, while rare, are real; zone-redundancy is strong protection, not immunity. Regions are the next circle out: fully separate geographies, and the unit that matters for law and latency.
For an Indian audience the decision is concrete. Residency: Indian regulators expect certain classes of data — payment data in particular — stored in India, which makes an Indian region a requirement rather than a preference for those workloads. Latency: a user in Hyderabad reaches a Mumbai or Hyderabad region in a few tens of milliseconds and a distant continent in several times that — for interactive applications the nearby region wins on feel alone. All three major providers run Indian regions.
Edge locations complete the picture: hundreds of small sites worldwide serving cached copies of your static content. Your application runs in one or two regions; your images and pages can still arrive from a rack near the reader. The division of labour — application in a region, static content at the edge — is the standard shape of a fast site.
What you should now be able to explain or do
Choose a region for a stated workload and justify it on latency, law and cost in three sentences. Say what an availability zone protects against and what it does not. Say when an edge location helps and when only a nearby region will do.
Check yourself
What does spreading an application across two availability zones protect against?
The failure of one zone — its power, cooling or networking. It does not protect against a region-wide incident or a regional control-system failure.
Why might an Indian payments workload be forced into an Indian region?
Data residency: Indian regulators expect payment data stored in India, so the region choice is made by law before latency or cost get a vote.
A site's images load from an edge location. Where does its application logic run?
In a region — edge locations serve cached static content near the user; the application itself lives in the region you chose.
What three inputs decide a region choice?
Latency to your actual users, the law that applies to your data, and the cost of that region.
Go deeper
We haven't checked most of these for screen reader use yet.
Back to Regions, availability zones and edge locations: work through the checklist