The Datomic Architecture and Data Model
summary
Datomic presents a concrete architecture: facts as entity-attribute-value-time records, serialized transactions handled by a single transactor, and application peers that query local in-memory indexes plus immutable storage segments in a key/value service. It implements datalog queries in-process, transaction functions that expand to assertions/retractions, a two‑tier cache and durable persistent trees for live merge and GC. Reads are immutable snapshots (database-as-value) and writes are transactional, giving scalable, consistent writes and elastic, parallel query capacity—an implemented system that works in production.
AI-written summary · about
chapters (15)
- 0:00Datomic introduction and goals
- 3:38Traditional DB problems
- 8:57Core model and process reification
- 12:50Read/write separation for scale
- 15:39Facts as EAVT triples
- 18:49Architecture breakdown
- 23:47Immutable trees and storage
- 27:14DB as value and caching
- 30:03Transaction functions
- 32:31Datalog programming
- 34:58Consistency, scale, query elasticity
- 37:50Schema and entity orientation
- 39:11Time as first-class
- 42:20Model and implementation
- 43:47Q&A on deletion
related talks
51mDatomic
40mThe power of Datomic database
40mDatomic: The Most Innovative DB You've Never Heard Of
37m11 insights after 11 years with the functional database Datomic
35mIt's Just Data
37m11 insights after 11 years with the functional database Datomic
52mEnhancing Datalog for Datomic
40mBuilding a Data Platform with Datomic