About Combining Functional Reactive Programming and Replicated Data Types
summary
Claim: combining FRP with CRDT-style replicated data types yields a practical, composable programming model that keeps sequential code and gives eventual consistency. Implementation details: counters store a per-replica map and values sum across replicas, lists use a replicated list with metadata, merges come from a lattice interface and merge function generation for primitives, maps and case classes. Built pieces: forwarding/replicated forwards, summarization to bound state, encryption-as-an-RDT; prototypes (time-tracking UI, student projects) run with good performance but removals and some data-model choices require careful design.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:45Collaboration
- 2:33SRT
- 6:30Decentralization
- 9:58Merge
- 14:35Replicate
- 18:13More Documentation
- 18:34Questions
- 19:02Replicated Data Types
related talks
36mConflict-Free Replicated Datatypes (CRDT)
29mWhat is Functional Reactive Programming?
39mEventually Consistent Data Structures
46mFunctional Reactive Programming
48mCRDTs: From sequential to concurrent executions
29mLet's replicate state together
28mNew insights into distributed and concurrent programming
31mFunctional && Reactive