Vždyť je to hned to první, co na tebe na google vyskočí:
"A failover cluster is a group of servers that work together to maintain high availability of applications and services. If one of the servers, or nodes, fails, another node in the cluster can take over its workload without any downtime (this process is known as failover)."