The Design and Organization of Data Centers/Redundancy

Everything will eventually fail. Redundancy allows you to minimize the damage of a system failure.

Failure characteristics
Planned vs. unplanned

Total vs. partial

Frequency of failure

Length of partial failure or outage

Structure
Ladder

Mesh

N+1

Implementation
Active/active

Active/passive

Human Factors
Notification

Unattended problem resolution

Documentation and problem clarity

Allowance for no-impact maintenance