Amazon has a massive rapidly growing global network of data centers connected through high speed fiber optic cables. This infrastructure is divided into regions, availability zones (AZ), and edge locations.
A region is physical location somewhere in the world with multiple availability zones. Every region is physically isolated and independent of every other region in terms of location, power, and water supply. Not all services are available in all regions. All new services first become available in US-East first.
Availability Zones one or more discrete data centers. An Availability Zone is a datacenter owned and operated by AWS. AWS services run in Availability Zones. Each region has at least two Availability Zones. Availability Zones have letter identifiers. For example us-east-1a, us-east-1b. Distributing instances across multiple Availability Zones allows failover if one Availability Zone goes down for some reason. The latency between Availability Zones is 10ms.
Edge Location is a data center owned by a trusted partner of AWS. These locations server request for CloudFront and Route 53. Requests are routed automatically. S3 Transfer Acceleration traffic and API Gateway endpoint traffic also use AWS Edge Network. This improves low latency.
What is the benefit of having some many data centers and physical locations? Low latency and fault tolerance.
It is faster and more cost effective to service an online request coming from Switzerland in Europe, rather than from South America. The response time is faster because of the distance.
Suppose everyone in the world is using only one datacenter somewhere in California. What will happen if that data center is hit by an earthquake, suffers from power outage, or has to be shutdown due to a forest fire nearby. Then every service being provide by that data center will be shut down. By distributing datacenters, we ensure continuity of service and fault tolerance.