Cloud Fundamentals

Cloud Regions and Availability Zones Explained

Learn how cloud regions and availability zones work, how they are named, and how to choose them for latency, data residency and disaster recovery.

On this page
  1. What is a cloud region?
  2. What is an availability zone?
  3. How regions and zones are named
  4. Why availability zones matter for resilience
  5. Why region choice matters
  6. Disaster recovery needs more than one region
  7. Other footprints: Local Zones, Wavelength, Outposts and edge locations
  8. Common misconceptions
  9. Frequently asked questions

A cloud region is a separate geographic area where a provider runs its data centers. An availability zone (AZ) is one or more isolated data centers inside a region, each with its own power, cooling and networking. Regions and zones are how cloud providers let you decide where your resources run and how well they survive a failure.

What is a cloud region?

A region is a geographic area, such as a metropolitan area or a country, that contains several isolated groups of data centers. AWS, Microsoft Azure and Google Cloud each operate dozens of regions around the world.

When you create a resource, you usually choose a region. Your data and workloads stay in that region unless you configure them to move. A few services are global or span several regions, so check each service’s documentation.

What is an availability zone?

An availability zone is one or more discrete data centers with independent power, cooling and networking. Zones in the same region are far enough apart to reduce the chance that one local event affects them all, but close enough to communicate quickly. On AWS, zones are typically within about 100 km (60 miles) of each other, and network latency between them is usually in the single-digit milliseconds.

Most AWS regions have at least three zones. The exact number varies by region and provider.

How regions and zones are named

Naming differs by provider, which is a common source of confusion.

Provider Region example Zone example
AWS us-east-1 (N. Virginia) us-east-1a
Google Cloud us-central1 us-central1-a (called a zone)
Microsoft Azure East US Zones 1, 2 and 3 within East US

AWS maps zone letters such as “a” to physical zones separately for each account. Two accounts can both see “us-east-1a” but use different physical zones. AZ IDs, such as use1-az1, identify the same physical zone in every account, so use those when you need to coordinate across accounts.

Why availability zones matter for resilience

Running an application in one zone means a failure in that zone can take the application down. Running copies across two or more zones in the same region protects against many single-zone problems, such as a power or cooling failure in one facility.

Multi-zone design requires more than extra servers. You also need a load balancer that routes traffic to healthy copies, and usually auto scaling to replace failed instances. AWS’s published EC2 SLA is higher for instances spread across two or more zones than for a single instance. Check the current SLA for the services you use.

High availability is not automatic. A single instance in a single zone can still go down.

Why region choice matters

Choosing a region affects several things at once:

  • Latency: Light in fiber travels about 1 ms round trip per 100 km, before routing delays. A region close to your users generally responds faster.
  • Data residency: Rules such as the EU’s GDPR can affect where personal data may be stored and processed.
  • Service availability: Not every service launches in every region, and new services often reach some regions first.
  • Price: Costs differ by region, so the cheapest region is not always the closest one.

Disaster recovery needs more than one region

Deploying everything in one region does not protect you from a region-wide outage. A widely reported outage in AWS’s us-east-1 region in December 2021 disrupted many services, which shows why regional independence matters.

Data does not copy itself to another region. You have to configure cross-region backups or replication yourself. Azure pairs some regions, such as East US and West US, for replication and staggered updates, but not every region has a pair.

Other footprints: Local Zones, Wavelength, Outposts and edge locations

Providers also offer smaller footprints that are not full regions:

  • AWS Local Zones extend a region’s services closer to large population centers.
  • AWS Wavelength Zones place compute inside telecom 5G networks for low-latency applications.
  • AWS Outposts installs AWS hardware in your own data center.
  • Edge locations, such as those used by content delivery networks, cache content and run limited code close to users.

Edge locations are not regions, and they do not replace one. For a comparison of centralized cloud and processing near the data source, read cloud computing vs edge computing.

Common misconceptions

  • An availability zone is one data center. It is one or more data centers with independent infrastructure.
  • One region protects against any outage. It does not protect you from a region-wide failure.
  • Edge locations are regions. They are much smaller and mainly cache content.
  • The same zone name is the same physical zone everywhere. On AWS, zone letters are mapped per account, so use zone IDs when you need consistency.
  • Data replicates across regions automatically. It does not, unless you configure it.

Frequently asked questions

What is the difference between a region and an availability zone?

A region is a geographic area with several isolated groups of data centers. An availability zone is one of the isolated groups inside that region, with its own power, cooling and networking.

Should I always deploy in the region closest to my users?

Usually it is a good starting point for latency, but check that the services you need are available there, that your data rules allow it, and that the price fits your budget.

Do I need more than one availability zone?

For production workloads that must stay online through a single facility failure, usually yes. Spreading copies across zones is one of the simplest ways to improve availability, and it works best with a load balancer and auto scaling. For a broader view of how these services are built, see how cloud computing works.

The complete guide to cloud fundamentals

How Cloud Computing Works (With Diagram)
Read the guide

Keep exploring

Guides from other topics

All topics →