CRDTs, Automerge: Local-First Software; book: Designing Data-Intensive Applications

In the context of computer science and distributed systems, CRDTs (Conflict-free Replicated Data Types) are specialized data structures that allow multiple users or systems to update the same data independently and concurrently without a central coordinator. They are designed to automatically merge divergent copies back into a single consistent state, making them a foundation for local-first and real-time collaborative software

1. Conflict-free Replicated Data Types (Computing)

CRDTs ensure Strong Eventual Consistency (SEC). This means that while different users might see slightly different versions of data at the same moment, once all updates have been received, every copy is guaranteed to be identical without any manual conflict resolution. [1, 2, 3, 4]

Key Characteristics

• Decentralized: No central server is required to decide the "correct" version; every node is a peer.
• Offline Support: Users can make changes while disconnected; these are merged seamlessly upon reconnecting.
• Mathematical Properties: To work, their merge operations must be commutative (order doesn't matter), associative (grouping doesn't matter), and idempotent (duplicate updates don't change the result). [1, 2, 3, 4, 5, 6]

podcast

SE Radio 716: Martin Kleppmann Local-First Software – Software Engineering Radio

References

• Automerge.org Open source project
  Built for Javascript and Rust, with ports or bindings for Swift, Python, C, Java, and more.
  
  Library Initialization (docs)
  
  automerge/automerge: A JSON-like data structure (a CRDT) that can be modified concurrently by different users, and merged again automatically. @GitHub
  
  Automerge: Automerge is a library of data structures for building collaborative applications.
  
  Automerge is a Conflict-Free Replicated Data Type (CRDT), which allows concurrent changes on different devices to be merged automatically without requiring any central server.
  
  
• Designing Data-Intensive Applications (DDIA) — an O’Reilly book by Martin Kleppmann (The Wild Boar Book)
  
  
• The past, present and future of Local-first software. Local-first conference in Berlin (May 2024) The past, present, and future of local-first — Martin Kleppmann’s talks
  
  
• Local-first software and geopolitical risk Local-first software and geopolitical risk — Martin Kleppmann’s talks
  
  
• Very revealing post about the economics of publishing a very successful technical book Writing a book: is it worth it? — Martin Kleppmann’s blog
  
  
• Local-first software seminal essay Local-first software: You own your data, in spite of the cloud
  
  
• Local-first conf 2026
  
  
• Using automerge Practical Local-First Software with Automerge – Peter van Hardenberg
  
  
• Martin Kleppmann on Bluesky.
  
  

Amazon.com: Designing Data-Intensive Applications, 2nd Edition: The Big Ideas Behind Reliable, Scalable, and Maintainable Systems (Audible Audio Edition): Martin Kleppmann, Chris Riccomini, Graham Mack, Ascent Audio: Books

Amazon.com: Designing Data-Intensive Applications: The Big Ideas Behind Reliable, Scalable, and Maintainable Systems (Audible Audio Edition): Martin Kleppmann, Benjamin Lange, Ascent Audio: Books