The CAP theorem is a principle of distributed computing which states that a system storing data on several networked machines cannot guarantee all three of the following at once: consistency, meaning every read sees the most recent write; availability, meaning every request gets a response; and partition tolerance, meaning the system keeps working when network links between machines fail. Computer scientist Eric Brewer put it forward as a conjecture in 2000, and Seth Gilbert and Nancy Lynch published a formal proof in 2002.
Since network failures cannot be ruled out, designers in practice choose what to give up during a partition. Bank ledgers and payment systems usually favour consistency, refusing some requests rather than risking double spending. Public blockchains often favour availability, letting nodes keep producing blocks and resolving conflicts later through chain reorganisations, which is why their finality is often probabilistic. The theorem is closely related to the blockchain trilemma, though the two describe different trade-offs.
Ref 1300000280