Regions, Availability Zones and Subnet Scope

Failure domains shape your network

A region is a geographic area, such as Mumbai or Hyderabad, containing several Availability Zones (AZs). Each AZ is one or more data centres with independent power, cooling and networking, connected to the other AZs in the region by low-latency links. AZs exist so that a fire or power failure in one building does not take down your whole application. This directly affects subnet design. In AWS a subnet belongs to exactly one AZ, so a highly available tier needs one subnet per AZ, for example three app subnets across three AZs. In Azure a subnet spans the whole region and zone choice is made per resource. In Google Cloud a subnet is regional, covering all zones in that region. The practical rule is the same everywhere: spread each tier across at least two zones, and make sure anything that tier depends on (NAT, load balancers, databases) is also zone-redundant.

Branches of a bank

A bank in Mumbai has several branches in different neighbourhoods. If one branch floods, customers walk to another. A region is the city, AZs are the branches, and putting your servers in only one AZ is like keeping all your cash in a single branch.

Zonal dependencies hide in plain sight

Teams spread instances over three AZs and then route all of them through a single NAT gateway in one AZ. When that AZ fails, every private instance loses internet access. Check every dependency, not just compute.

त्वरित जाँच: In AWS, what is the scope of a single subnet?

• One Availability Zone
• The whole region
• All regions in the account
• One EC2 instance