Regions and Availability Zones · Chapter 1 of 1
How AWS physical locations affect latency and reliability
AWS partitions its global footprint into independent Regions and Availability Zones to let you trade speed, cost, and disaster recovery.
FoundationsBuilds the idea from nothing. No prior AWS assumed.
Regions and Availability Zones
What is the physical difference between an AWS Region and an Availability Zone?
Four instances behind a load balancer, split across two Availability Zones. Every one of them is healthy and taking traffic.
- 01
Regions are separate geographic areas
An AWS Region is a completely separate geographic area, such as Ireland or Singapore. Each Region contains multiple, isolated locations. Regions are independent of each other, meaning a power failure or natural disaster in one will not impact the operation of another.
- 02
Availability Zones are clusters of data centres
An Availability Zone is one or more physical data centres within a Region. They are separated by miles to ensure a fire, flood, or power grid failure in one does not affect another. However, they are close enough to be connected by private, low-latency fibre networks.
- 03
Selecting your home Region
You must choose a Region before creating resources. This choice is guided by three constraints: latency to your users, local data residency laws, and service pricing. Deploying a database close to your customers reduces network round-trip time from 150 milliseconds to under 10 milliseconds.
- Instances serving
- 0 of 0
- Status
- nothing deployed
- Latency to users
- 14 ms
- Relative price
- 1.08×
Check yourself
Which criteria is the most critical when data residency laws require customer data to remain within a specific country?
Under the hoodThe same thing from underneath: limits, failure modes, numbers.
The engineering cost of redundant zones
How do latency and data costs change when you distribute systems across multiple zones?
Four instances behind a load balancer, split across two Availability Zones. Every one of them is healthy and taking traffic.
- 01
Low latency fibre links
Availability Zones inside a Region are connected by redundant, private dark fibre networks. Because the physical distance between zones is typically under 100 kilometres, the round-trip network latency is extremely low. This allows applications to write to databases in different zones synchronously.
- 02
The hidden price of reliability
While copying data between zones is necessary for reliability, it is not free. AWS charges a fee for every gigabyte of data that crosses an Availability Zone boundary. If a web server in zone A reads from a database in zone B, you pay for the data transfer in both directions.
- 03
Designing for zone survival
A single Availability Zone failure is a standard event that AWS expects you to handle. By deploying load balancers and virtual machine groups across at least two zones, your system can survive the loss of an entire building. The traffic is redirected to the healthy zone within seconds.
The numbers
- Inter-AZ network latency
- Under 2 millisecondsEnables synchronous replication, such as database clustering, without degrading application performance.
- Cross-AZ data transfer charge
- $0.01 per GB in each directionSending 10 TB of data between servers in different zones costs $200 in total.
- Minimum Availability Zones per Region
- At least 3 zonesAWS guarantees this minimum for all modern Regions to allow quorum-based clustering.
Check yourself
A web application in Availability Zone us-east-1a queries a database in us-east-1b, transferring 100 GB of data. How is this billed?
Official references & further reading
These lessons simplify selected behaviors for learning. Verify current service limits, Region support and production requirements with the official references. Experiments describe their own assumptions.
Report an error or suggest a clearer explanation →