Home▸Distributed & Parallel Computing▸CRDTs: Merging Data Without a Central Coordinator

🔗 CRDT: Conflict-Free Replicated Data Types

Explore how CRDTs let distributed replicas be updated independently, offline, and out of order, then merge automatically into the same final state without a central coordinator.

Distributed & Parallel Computing3DModerate60 FPS

This simulator demonstrates how independent replicas of a CRDT, each updated offline and out of order, always converge to an identical final state once merged, by visualizing the join-semilattice merge operation on a live Grow-only Counter.

🔬 What It Demonstrates

🎮 How to Use

Add replicas, apply increments to each one independently while they remain disconnected, then trigger merges in any order or sequence you like and watch every replica's total converge to the same value regardless of merge order.

💡 Did You Know?

Did you know that the mathematical property guaranteeing CRDT convergence, the join-semilattice, is the same algebraic structure used to define the least upper bound in order theory, meaning CRDT correctness proofs borrow directly from decades-old pure mathematics rather than being a new invention of distributed computing.

⚙ Under the hood

Explore conflict-free replicated data types, distributed data structures that let independently updated replicas always merge into the same final state.

crdtdistributed-systemsreplicationeventual-consistencyg-counterlocal-first-softwarecomputer-sciencedatabases

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install