Eric Brewer

Author of the CAP theorem (2000 PODC keynote) and its 2012 follow-up that revised the “pick two” framing. UC Berkeley professor and VP of Infrastructure at Google.

Why he matters here

• 2000: Introduced the cap-theorem|CAP theorem at PODC, formalized by Gilbert and Lynch in 2002.
• 2012: Published a follow-up paper calling the “pick two” framing misleading — not wrong, incomplete. The paper argues:
  1. C, A, and P are all spectrums, not booleans.
  2. CAP only describes partition behavior; it’s silent on what to optimize the rest of the time (the gap pacelc addresses).
  3. Systems should design for three modes (normal, partition, recovery) deliberately — see partition-mode-design.

The 2012 paper also brought stand-in mode (the ATM pattern) and partition compensation into the canonical CAP vocabulary, connecting the discussion to the saga-pattern from Garcia-Molina & Salem (1987).

See also

• cap-theorem — the theorem and the 2012 revision
• partition-mode-design — Brewer’s three-state design pattern
• pacelc — Daniel Abadi’s extension that addresses what CAP omits