Cloud Design Patterns: Event-Driven, CDC, and Mesh Architectures

These patterns power the modernization projects I run with customers. The cdp-mesh-tutorial is the Polytech Montpellier course material — hands-on Kafka and CDC exercises on OpenShift, from zero to a running event-driven pipeline in a few hours.

The most impactful pattern for IBM i: Debezium CDC. Capture every Db2 for i journal record change and stream it to Kafka — zero polling, no ETL, real-time data propagation without touching the application. The debezium-ibmi-demo repo shows this end-to-end: IBM i journal → Debezium → Kafka → downstream consumers.

Data mesh is the next step: treating IBM i operational data as a domain-owned, catalogued data product, accessible to analytics and AI workloads via federated query — without copying it into a warehouse.