Failover is the automatic switch to a redundant standby system when the primary system fails. The goal is minimal or zero service interruption.
Failover mechanisms operate at every level: network (redundant links, HSRP/VRRP), server (clusters with hot standby), storage (RAID, replicated databases), and application (load balancers with health checks). Failover time directly affects the RTO (Recovery Time Objective) in the disaster recovery plan. Automated failover requires regular testing — a failover that does not work in an actual emergency is worthless.