A Conflict-free Replicated Data Type, or CRDT, is a data structure built so that multiple replicas can be modified independently, even while disconnected, and later merged with no coordination and no possibility of conflict. The term and the unifying theory were introduced by Marc Shapiro, Nuno Preguica, Carlos Baquero, and Marek Zawirski in their 2011 work “Conflict-free Replicated Data Types,” published as INRIA research report RR-7687 and at the SSS 2011 symposium.
The key idea is to constrain the data type’s operations so that concurrent updates always commute or merge deterministically. Shapiro and colleagues describe two complementary styles: state-based CRDTs, which ship their whole state and merge it using a least-upper-bound operation defined over a lattice, and operation-based CRDTs, which broadcast individual operations that are designed to commute. Either way, any two replicas that have seen the same set of updates are guaranteed to reach the same state.
The paper formalizes this property as Strong Eventual Consistency: replicas that have delivered the same updates have equivalent state immediately, without the conflict-resolution rollbacks or last-writer-wins data loss that earlier eventually consistent systems relied on. The authors give concrete designs for counters, sets, graphs, and other structures, showing the approach is practical rather than purely theoretical.
CRDTs became the foundation of offline-first and real-time collaborative software. Because edits made on different devices merge automatically and converge, they underpin collaborative text editors and shared documents, where many people type at once and every replica must end up showing the same result.