What is an availability zone
In the context of cloud computing, it is a public cloud provider’s data center that contains its own power and network connectivity. There are typically multiple Availability Zones in a region. Each region is a separate geographic area, and each region generally has multiple, isolated locations known as Availability Zones. A common misconception is that a single zone equals a single data center. In fact, each zone is usually backed by one or more physical data centers, with the largest backed by as many as five. While a single availability zone can span multiple data centers, no two zones can share a data center. Abstracting things further, to distribute resources evenly across the zones in a given region, some providers independently map zones to identifiers for each account. This means an east coast availability zone for one account may not be backed by the same data centers as an east coast zone for another account.
How availability zones may be implemented
- In each zone, participating data centers are usually connected to each other over redundant low-latency private network links.
- All zones in a region communicate with each other over redundant private network links.
- These intra and inter-zone links are heavily used for data replication by a number of cloud provider services including storage and managed databases.
Availability zones benefits
- Lower latency where multiple availability zones are implemented, it makes a sense to have the servers that provide a given application located relatively close to the end users that will access that application. Latency is a big issue in the application world, and many cloud providers address is this by distributing servers and storage, placing those resources closer to their customers’ end users.
- Customers can deploy their applications and instances across availability zones, and design their environment so that if one availability zone fails, the instances in the other availability zone will become active, and continue the work of the servers in the failed availability zone, until such time as service can be restored.