AWS infrastructure is spread across the globe to provide low latency and high availability to customers all around the world. The components of AWS global infrastructure are
- Regions
- Availability Zones
- Edge Locations
- Regional Edge Caches
Region
An AWS region is a geographical area that is completely independent from other AWS regions. Each region is connected to 2 independent and completely redundant transit centers that connect regions together. Each AWS region has multiple partitions known as Availability Zones. Not all regions have the same number of availability zones, eg: US East (N. Virginia) has 6 and Canada (Central) has 2 availability zones.
AWS does not automatically replicate data across regions thus making it easy for the customers to comply with regional requirements. Not all services are offered in all the AWS regions. Here is the aws page showing services offered at various regions. Several factors must be considered while choosing an AWS region for your application including regional compliance requirements, AWS services offered and proximity to end users.
Availability Zone
An availability zone is a cluster of datacenters that is fully isolated from other availability zones. Each availability zone has it’s own power infrastructure. Availability zones are built several miles apart to achieve complete isolation and fault tolerance.
Each datacenter is associated with one and only one availability zone and each availability zone is associated with one region. Availability zones within a region are interconnected by fully redundant metro fibers to provide high throughput and low latency networking between them. Availability zones play an important role in designing highly available and fault tolerant applications.
Edge Location
An edge location is an AWS facility used to cache data and reduce latency for end users. These are deployed in populous cities all around the world. These Points of Presence are used by Amazon CloudFront (Amazon’s content delivery network(CDN)) to deliver content to end users with low latency.
Regional Edge Cache
Regional edge caches are facilities that are placed between availability zones and edge locations. These facilities have larger cache-width than edge locations. Thus when the content expires at edge locations, cloud front can obtain it from regional edge cache instead of going back to the origin thus reducing latency and improving performance.
Currently there are 20 regions and 61 Availability Zones across the globe.
- US East (N. Virginia)(6) us-east-1
- US East (Ohio)(3) us-east-2
- US West (N. California)(3) us-west-1
- US West (Oregon)(4) us-west-2
- Asia Pacific (Mumbai)(2) ap-south-1
- Asia Pacific (Seoul)(2) ap-northeast-2
- Asia Pacific (Singapore)(3) ap-southeast-1
- Asia Pacific (Sydney)(3) ap-southeast-2
- Asia Pacific (Tokyo)(4) ap-northeast-1
- Asia Pacific (Osaka-Local)(1) ap-northeast-3
- Canada (Central)(2) ca-central-1
- China (Beijing)(2) cn-north-1
- China (Ningxia)(3) cn-northwest-1
- Europe (Frankfurt)(3) eu-central-1
- Europe (Ireland)(3) eu-west-1
- Europe (London)(3) eu-west-2
- Europe (Paris)(3) eu-west-3
- Europe (Stockholm)(3) eu-north-1
- GovCloud US (US-East)(3) us-gov-east-1
- GovCloud US (US-West)(3) us-gov-west-1
- South America (Sao Paulo) sa-east-1
Below 4 new regions have been announced with a total of 12 Availability zones
- Bahrain
- Cape Town
- Hong Kong SAR
- Milan