A distributed commit log is a replicated, append-oriented record of events maintained across multiple nodes. Kafka stores messages as an immutable event sequence rather than treating them only as transient broker traffic. section 1.7.3
This design shifts integration from point-to-point data movement toward a central, durable, replayable stream. Persistence, replication, and historical retention support high-throughput workloads, fault tolerance, recovery, and event-sourcing use cases. section 1.7.3 section 12.0
Architectural consequences include:
- Consumers can recover data after downtime instead of depending solely on an in-flight message exchange. section 12.0
- Historical events can be retained and reprocessed, making the log suitable for event-sourcing scenarios. section 1.7.3
- Event modeling becomes an architectural concern: teams identify sources, define captured events, and establish schemas for compatibility and scalability. section 1.7.3
Enterprise event foundation
A customer-data architecture can capture transactions, customer interactions, and social-media mentions as events, then expose them to connectors and real-time stream-processing applications through Kafka’s durable log. section 1.7.3
