A fundamental concept of AWS is that its services and the solutions built on top of them are architected for failure. This means that a failure of the underlying resources is a scenario actively planned for, rather than avoided until it cannot be ignored.
Due to this, all the services and resources available are divided up in to geographically diverse Regions. Using specific regions means you can provide services to your users that are optimized for speed and performance.
Within a region, there are always multiple Availability Zones (a.k.a. AZ). Each AZ represents a geographically distinct—but still close—physical data center. AZs have their own facilities and power source, so an event that might take a single AZ offline is unlikely to affect the other AZs in the region.
The smaller regions have at least two AZs, and the largest has five.
At the time of writing, the following regions are active:
| Code | Name | Availability Zones |
| us-east-1 | N. Virginia | 5 |
| us-east-2 | Ohio | 3 |
| us-west-1 | N. California | 3 |
| us-west-2 | Oregon | 3 |
| ca-central-1 | Canada | 2 |
| eu-west-1 | Ireland | 3 |
| eu-west-2 | London | 2 |
| eu-central-1 | Frankfurt | 2 |
| ap-northeast-1 | Tokyo | 3 |
| ap-northeast-2 | Seoul | 2 |
| ap-southeast-1 | Singapore | 2 |
| ap-southeast-2 | Sydney | 3 |
| ap-south-1 | Mumbai | 2 |
| sa-east-1 | Sao Paulo | 3 |