Datomic at Nubank, Simplicity Scaled
summary
Nubank runs Clojure and Datomic at 114.3 million users, handling petabytes, 3,000 Datomic databases, 4,000 microservices, 72 billion daily Kafka events and millions of requests per second. Datomic uses a peer model with a single active transactor that peers read from storage and cache locally. Newer features like transactor hints and implicit partitions let the transactor prefetch IO and improve data locality, measurably reducing IO, lowering latency and cutting cost to serve.
AI-written summary · about
chapters (10)
- 0:00Datomic at Nubank
- 1:03Bank growth in Brazil
- 5:19Scalability in modern systems
- 8:50Datomic peer model
- 12:41Datomic transactor and storage
- 15:30Transactor performance hints
- 17:36Data locality in Datomic
- 20:08Simplicity in banking
- 21:54Team scalability and focus
- 24:25Datomic resources
related talks
26m"Operating Datomic at Scale"
51mHow Nubank leverages Datomic's database as an immutable value to run at scale
51mDatomic
40mDatomic: The Most Innovative DB You've Never Heard Of
40mThe power of Datomic database
46mThe Datomic Architecture and Data Model
37m11 insights after 11 years with the functional database Datomic
38mDatomic Made Easy Datomic in the Cloud