How we scaled a payments system with MongoDB
summary
Nets migrates the Easy payment platform from an RDBMS to MongoDB Atlas, embedding most payment data as single documents to avoid joins and using sharding across Europe for low-latency, GDPR-compliant storage. They apply subset pattern for high-cardinality actions, preserve transactional ACID semantics, use indexes and aggregation pipelines instead of heavy SQL views, and offload cold data to object storage. The result scales from tens of millions to billions of transactions per month with lower cost and acceptable operational complexity.
AI-written summary · about
chapters (7)
- 0:00Start
- 12:04What is a Document Database?
- 15:43Transactional Writes
- 16:35Native Object based programming
- 17:29Cross Data Centre HA
- 29:14Indexes in MongoDB
- 31:49Aggregation Pipelines
related talks
18mFrom SQL to NOSQL - a gentle intro for Devs
25mHybrid Applications with MongoDB and RDBMS
43mHelp! I Need To UnSQLize My Application
28mDatabases Can't Be Just Databases Anymore
29mNoSQL to Distributed SQL: Best Practices and Lessons Learned by Helpshift
29mModern Database Architecture: From SQL to NoSQL & Distributed SQL
01h 01mFrom SQL to Azure Cosmos DB
35mPainless Data Storage with MongoDB and Go