Deploy redundant hardware everywhere
“Shared” storage accessible by all nodes
Clustered application writes data to shared storage so all nodes can assess it
Registry settings containing current cluster configuration information is replicated to each node
Nodes monitor the health of other nodes