The AWS global infrastructure primarily consists of Regions and Availability Zones.
In addition to these core divisions, there are Edge Locations, commonly known as Points of Presence, utilized by services like CloudFront, Route 53, and Global Accelerator.
Furthermore, AWS includes Local Zones, which extend the reach of Regions to additional physical locations.
Additionally, there are Wavelength Zones, specifically designed to cater to 5G networking needs.
Finally, the AWS global infrastructure encompasses any data center or server that employs the AWS Outpost Family services
AWS Regions
An AWS Region refers to a specific geographical area where AWS operates several Availability Zones. These Regions comprise multiple, separate Availability Zones linked with low-latency connections.
Examples of regions are: Northern Virgina (USA), Ohio (USA), Paris (Europe).
Regional services
AWS Regions allow users to strategically locate resources, including EC2 instances, RDS databases, DynamoDB tables, and others, closer to their end users, thereby boosting performance.
Typically, resources configured in one Region are isolated and not accessible from other Regions by default.
Utilizing regional service endpoints requires adherence to a specific regional notation format. For instance, when accessing an EC2 instance in the US West (Oregon) Region, the endpoint would be formatted as ec2.us-west-2.amazonaws.com. This notation combines the service name (ec2), the region (us-west-2), and the domain (amazonaws.com), ensuring precise regional targeting and access.
Global services
Certain AWS services, including AWS IAM, AWS Organizations, Amazon CloudFront, AWS Firewall Manager, AWS Global Accelerator, and AWS Direct Connect, are global in nature.
This means they are not confined to a specific region, and it's neither necessary nor possible to assign them to any particular region, for various reasons depending on the service
For instance, it's essential for users, roles, or policies in AWS IAM to be accessible across all regions. Similarly, services like Route 53 are designed to operate as a single instance to effectively manage and route traffic.
Quasi global services
Some AWS services, such as Amazon S3, exhibit both global and regional characteristics. For instance, S3 is managed and accessed globally, but it functions at a regional level regarding data storage.
S3 employs a global namespace, which requires each bucket to have a unique name across the entire AWS platform, regardless of the region.
AWS Availability Zones and High Availability
An AWS Availability Zone (AZ) comprises one or more discrete, individually contained data centers, each equipped with redundant power and networking.
Situated within an AWS Region that always contains multiple AZs, each AZ is physically isolated to protect against power failures, network disruptions, and similar issues.
High Availability
The term High Availability is closely associated with Availability Zones (AZs) in AWS, implying that a service is highly likely to remain available even in the event of issues or even a disaster. This is due to the replication of the service across multiple AZs, underscoring the concept of an 'Availability Zone'.
Natively High available services
Some services are replicating automatically to many AZs by default. For example Amazon S3, Amazon DynamoDB, Amazon ElastiCache, Amazon Route 53 and AWS Lambda.
Services supporting HA
While other services may not be highly available by default, they can achieve this status through appropriate configuration. For instance, Amazon EC2 and Amazon RDS can be configured for high availability
Edge Locations
This group forms the largest segment of AWS's global infrastructure, with more than 600 Edge Locations also refered as Points of Presence (PoPs) worldwide. AWS clarifies that these are neither traditional data centers nor standalone servers but not specify further.
So, what exactly are they? It's reasonable to speculate that they comprise server setups for content caching, along with networking equipment, load balancers, and similar hardware. This configuration is likely designed to extend AWS's network reach, particularly for services like CloudFront, Route 53, and Global Accelerator
Local Zones
Local Zones are typically positioned closer to users than Regions and their Availability Zones. They are particularly beneficial for latency-sensitive applications or where adherence to specific local data residency laws is necessary.
Each Local Zone maintains a direct connection to a parent Region. This setup allows for services that need low latency or specific compliance requirements to be physically deployed within the zone, while other services can be accessed through their APIs.
Nevertheless, there might be some restrictions in accessing the entire array of AWS services from the parent Region, due to a variety of reasons.
Commonly utilized services in Local Zones include EC2, EBS, EFS, RDS, ECS/EKS, and Systems Manager.
Wavelength Zones
As a relatively new addition to AWS infrastructure, Wavelength Zones are akin to Local Zones, primarily serving latency-sensitive applications. However, their focus is on 5G networks. Wavelength Zones are strategically situated near mobile network operators' data centers and maintain a connection to a parent AWS Region, similar to how Local Zones operate.
There are several restrictions regarding the services that can be deployed in Wavelength Zones. Currently, it's established that services such as EC2 instances and associated offerings like VPC, subnets, EBS, and others are supported.
Additionally, only specific types of EC2 instances are compatible with Wavelength Zones, which further narrows the scope of use cases for these zones.
AWS Outposts Family
A range of various servers and racks, deployable on-premises or at edge locations within the client's infrastructure, forms the foundation of a hybrid cloud setup. These servers allow numerous AWS services to be run locally instead of in AWS.
Key applications include latency-sensitive tasks, data residency requirements, challenges or limitations in migrating data to the public cloud, and the need for local data processing
Summary
| Component | Numbers | Description |
| Region | 33 | A geographic area with multiple isolated locations known as Availability Zones, hosting AWS's cloud infrastructure |
| Availability Zone | 105 | A single or group of data centers within a Region, providing redundant power, networking, and connectivity for high availability |
| Edge Location | 600+ | A site that AWS uses to cache content closer to users for faster delivery, primarily used in services like Amazon CloudFront and Route 53 |
| Local Zone | 36 | An extension of an AWS Region located closer to users, providing select services with low-latency access in specific geographic areas. |
| Wavelength Zone | 29 | An infrastructure deployment that embeds AWS services within telecommunications providers' data centers, enabling ultra-low latency applications on 5G networks |
| AWS Outposts Family | N/A | A fully managed on-premises infrastructure and services from AWS, designed to provide a consistent hybrid cloud experience |
References
AWS Region - AWS SDKs and Tools
What Is Amazon DocumentDB (with MongoDB Compatibility) - Amazon DocumentDB
AWS Region selection - AWS SDK for Kotlin
Static stability using Availability Zones
Availability Zones - AWS Fault Isolation Boundaries
FAQs | What is Amazon Cloudfront CDN?
Security at the Edge: Core Principles - Security at the Edge: Core Principles
What is AWS Outposts? - AWS Outposts
What is AWS Local Zones? - AWS Local Zones
AWS Local Zones FAQs - Amazon Web Services
What is AWS Wavelength? - AWS Wavelength
Available Wavelength Zones - AWS Wavelength
What is AWS Outposts? - AWS Outposts
What is AWS Outposts? - AWS Outposts
On-Premises Private Cloud - AWS Outposts Family - AWS