Building Enterprise Cache Based on CQRS - Komes Subramaniam, T-Mobile
summary
Describes a production CQRS read cache that denormalizes event streams into a presentation-friendly store using Spring Boot, JMS (TIBCO then RabbitMQ) and Cassandra, built in a three- to four-week pilot for order-status and subscriber views. Initial seeding uses login-triggered events, live handlers enrich and aggregate reference data, and an automated real-time comparator validates data against legacy systems. Result: APIs hit 99.9th-percentile latencies around 32 ms, offloads reads and cuts operating cost.
AI-written summary · about
chapters (10)
- 0:00Introduction and speaker background
- 7:17Performance challenges with legacy systems
- 11:10CQRS pattern overview and benefits
- 15:07First use case order status implementation
- 21:37Second use case subscriber details
- 26:15Results and benefits achieved
- 26:57Challenges and solutions in implementation
- 30:13Building confidence in the solution
- 33:43Guidelines and best practices
- 34:32Q&A session
related talks
53mLeveraging Caching For Cloud Native Applications
46mMicroservices Data Patterns: CQRS & Event Sourcing
01h 30mCassandra, Couchbase and Spring Data in the Enterprise
51mMicroservices Data Patterns: CQRS & Event Sourcing
43mCaching and Beyond - Smarter Data Processing with Java
44mFrom REST to CQRS with Clojure, Kafka, & Datomic
32mA Change-Data-Capture use-case: designing an evergreen cache
30mCaching patterns for lazy devs for lazy loading