Exploring four hidden superpowers of Datomic
summary
Shows concrete Datomic techniques used at Nubank: attach metadata to transactions to build a forensic audit trail and answer who/when/why; generate Last-Modified/If-Modified-Since by taking the max transaction time touching a customer’s owned entities; implement mobile sync via as-of and since filters to return only datoms after a given t; express ownership with reusable Datalog rules and enforce it by passing a filtered immutable DB; use with for projected future charges, virtual DBs for tests, single transactor on DynamoDB (5–7s failover, ~1–2GB data).
AI-written summary · about
chapters (12)
- 0:00Start
- 0:01Exploring Four Datomic Superpowers
- 0:19Why build a bank from scratch in Brazil using Clojure and Datomic?
- 3:54Audit Trail
- 7:42How long did the customer spend on each stage of the acquisition funnel?
- 11:03Authorization
- 20:46HTTP Cache
- 27:01Bonus
- 27:15Future DBS
- 29:16Testing
- 30:46Schema Extension
- 31:52Sharding Reads
related talks
40mThe power of Datomic database
37m11 insights after 11 years with the functional database Datomic
51mHow Nubank leverages Datomic's database as an immutable value to run at scale
37m11 insights after 11 years with the functional database Datomic
51mDatomic
46mThe Datomic Architecture and Data Model
41mReal-World Datomic: An Experience Report
52mEnhancing Datalog for Datomic