A design principle where systems have one more unit than strictly necessary, ensuring operation continues if any single component fails. "2N" redundancy doubles everything. Data centers apply redundancy to power feeds, cooling systems, network connections, and other critical infrastructure. Higher redundancy increases costs but improves reliability.
Discussed in Chapter 3 of This Is Server Country