Eventually Consistent Data Structures
summary
Explains why Riak, Voldemort and Cassandra need deterministic data types and presents Convergent/Commutative Replicated Data Types that guarantee a single merge result. Covers concrete implementations: G-counter and PN-counter for increments, G-set, 2P-set, LWW-element-set and OR-set (observed-remove) for membership, and graph constraints. Shows how CRDTs fix Dynamo-style cart and support social graphs, notes GC/tombstone growth and lists client libraries like statebox and knockbox.
AI-written summary · about
chapters (8)
- 0:00Introduction and setup
- 0:49Eventual consistency basics
- 8:03Conflict-free replicated data types
- 14:36CRDT flavors and registers
- 16:45Counters and sets
- 23:57Breather and graph data types
- 30:43Use cases and challenges
- 35:23Q&A session
related talks
36mConflict-Free Replicated Datatypes (CRDT)
48mCRDTs: From sequential to concurrent executions
40mPractical Demystification of CRDT
30mCONSCIOUS CONSISTENCY WITH AKKA CLUSTER, CRDTS AND DISTRIBUTED DATA
29mAbout Combining Functional Reactive Programming and Replicated Data Types
36mCRDT Architecture and Group Systems
52mThe Computer Science behind a modern distributed data store
34mUsing CRDTs Beyond Text Editors