Redundancy

The duplication of critical components or functions of a system with the intention of increasing reliability of the system.

Redundancy is the practice of duplicating critical system components so that if one fails, another can take over without service interruption. Redundancy can apply to hardware (multiple servers, power supplies, network links), software (multiple instances of an application), and data (replicated databases, backup copies).

Types of redundancy include active-active (all copies handle traffic simultaneously), active-passive (standby copies activate only on failure), and N+1 (one spare for every N active components). Geographic redundancy distributes copies across different physical locations or data centers to protect against regional failures.

In API infrastructure, redundancy is essential for maintaining availability. API gateways deployed across multiple regions or edge locations provide geographic redundancy, ensuring that if one location fails, requests are automatically routed to the nearest healthy location. Backend services behind the gateway should also be redundant. Serverless platforms provide built-in redundancy by running functions across multiple availability zones or edge nodes, removing the need for teams to manage redundancy configurations manually.