Region

diagram
The regions are all connected
- The regions are connected by a network owned and operated by AWS
- AWS owns and operates their own cables.
- As a customer, this gives secure, fast connectivity between the workloads
Consists of multiple data centers, each region is physically isolated from each other
For many services you can choose in which region it should run
You can rent multiple servers for reliability/split workload
Choosing the right region
- Different Pricing
  - different services have different costs depending on the Region
- Service Availability
  - Some services are only available in certain Regions
- Legal regions
  - compliance with data governance and legal req
  - ex) a company must store user data in the EU
- Availability & latency
  - proximity to your customers - the closer, the faster (reduce latency, improved performance)
  - workloads can be executed in multiple Regions to increase availability and reliability

Availability Zone (AZ)

Data centers in a Region are grouped into AZs, which are also separated from each other
A region will generally contain 3 AZs
One AZ contains at least 1 data center
Reliability & reliability: Each AZ is designed as an independent failure zone
- If an app is partitioned across multiple AZs, companies are better isolated & protected from issues like power outages, tornados, earthquake, etc

diagram
Smaller data centers distributed all over the world
- exists in addition to the data centers in the regions & AZs
they can receive data (ex. a copy of your website) and cache it
AWS makes sure that the requests to your website automatically reach the Edge location closest to a given customer so that latency stays low
Edge locations are optimized to handle traffic even if they aren’t the absolute closest in distance.
The goal is to reduce load on the main data center and serve cached content quickly
They’re part of a CDN (Content Delivery Network), like AWS CloudFront.

Smaller AWS regions which are close/inside metropolitan areas
For workloads that are super close to your customers… closer than edge locations
Perfect for ultra low-latency
Only a limited set of supported services (but has the main services)
You can extend VPCs from your regions to your local zones, so that services running in a local zone will be part of main region
- (ex. EC2 instances in a local zone will be part of the same VPC that also contains EC2 instances in a main AWS region)

kinda similar to local.. but not
Server racks, which you can add to on-premise infrastructure
- On-premise = to infrastructure that is physically located in your own data centers
Adding AWS managed infrastructure into your own data centers
You can order these racks
You can extend VPCs from your regions to your outposts, so that services running in a local zone will be part of main region
Comes in 3 form factors: 42U, IU, 2U
- 42U
  - A full rack of servers provided by AWS
- IU, 2U
  - servers that you can place in your existing racks (2U is bigger)
Allows hybrid environment
- You still have certain workload on your own machines, but you can combine that with AWS based services

special AWS infrastructure that are integrated with telecom networks and placed closer to where the data is generated, within 5G networks
- super small data centers running in the 5G networks
Ideal for scenarios where real-time data processing is needed at the edge of the 5G network
- achieves lowest latency possible
limited set of supported services
can connect services in wavelength zones → AWs region
claude:
- Telecom providers like Verizon or Vodafone have their own specialized data centers that form the backbone of their mobile networks. These facilities contain the networking equipment, servers, and systems that process all the mobile data traffic.
- Wavelength Zones are AWS compute and storage infrastructure physically installed inside these telecom data centers.
- By placing AWS resources directly within the telecom provider’s network infrastructure (rather than in separate AWS facilities), data doesn’t need to travel across the public internet or even between separate facilities to reach AWS services.