Redundancy
Having backups of critical systems so if one fails, another is ready to take over.
Think of it like
Having two parachutes instead of one.
Example
Three inference servers instead of one — any two going down still leaves one answering requests.
How it actually works
Redundancy costs money but buys uptime. For stateless services (like inference), it's simple — just run more copies. For stateful services, you need distributed consensus and careful coordination. Redundancy has to cover everything: servers, storage, network.
For product teams
Redundancy is the most straightforward way to improve uptime and user trust.
For engineers
Redundancy is cheap compared to downtime; over-invest here.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome