Scalable Architecture for Processing UPI Real-Time Transactions Using Distributed SQL
summary
Platform processes 4.6 billion UPI transactions in Feb 2022 and projects ~1 billion transactions per day, so the stack moves from monolithic DBs to microservices plus separated OLTP/OLAP with a Kafka connector for downstream processing. Goal is sub‑second responses with ~99.19% availability and a 99%+ success rate, deployed cloud or on‑prem and scaled across regions with YugabyteDB as the distributed SQL layer, active-active multi-site deployment, horizontal sharding, and automated failover and autoscaling.
AI-written summary · about
chapters (9)
- 0:00UPI and its growth
- 3:10UPI volume and adoption
- 8:46Mindgate’s UPI processing
- 10:11Monolith to microservices
- 13:01Six scalable design principles
- 16:56Legacy database challenges
- 19:43Distributed SQL for scalability
- 21:07Distributed SQL architecture
- 22:52Summary and future
related talks
20mPowering the Future of Payments with
18mBuilding a Scalable, Resilient System for Real-Time Fraud Detection
18mTowards a Lightning Fast System: Migrating a Real-Time Payment Service to
25mHow Sharekhan Scaled Trading Platform with Yugabyte's Distributed SQL
17mBuilding a Scalable and Performant SaaS Platform with YugabyteDB
11mChallenges Deploying Modern Real-Time Payment Systems on Traditional RDBMS
35mScaling UPI for the Future: Zero Downtime Migration to
30mFrom monolith to petabyte scale: Built a cloud-native, AI-ready data layer