Redundancy vs Failover
People think having a spare means the service stays up, and it does not on its own. Redundancy is owning the spare. Failover is the switch that moves the work to it when the main part dies. A spare that nothing switches to sits idle while the site is down, so the switch has to be built and tested too.
Redundancy (Systems)
A spare of each key part, so one failure does not stop the system.
Redundancy means keeping more than one of each important part, so one can fail and the system still works. A server may have two power supplies and two network links. A service may run on three servers when two could carry the load. That setup is called N+1: the number you need, plus one spare. Redundancy costs money every day, to save you on the day something breaks.
Failover
Switching work to a standby copy when the main one fails.
Failover is switching work from a part that failed to a standby that is ready to take over. In an active-passive setup, one copy does the work and the other waits. In an active-active setup, both share the work. Either one can carry it all. Failover can start on its own after a failed health check, or be done by hand. Moving back to the first copy once it is fixed is called failback.
In one line
- Redundancy (Systems)
- A spare of each key part, so one failure does not stop the system.
- Failover
- Switching work to a standby copy when the main one fails.