Mirroring tables from databases such as Postgres, MySQL or Oracle into a data lake makes transaction data broadly available for analytics while maintaining isolation for transactional databases.
In this Webinar, Jason Reid, Head of Product, Tabular, and Cliff Gilmore, Principal Solutions Architect, Tabular, discuss:
– Why CDC is technically challenging, including the need to create workload isolation, ensure strong consistency, and handle schema evolution
– Iceberg techniques to address common CDC challenges include using append-only change logs, continuous processing, MERGE patterns, and delta files to avoid write amplification
– A live CDC to Iceberg demo combining Debezium, the Kafka Connect Iceberg sink, and Tabular’s serverless CDC merge capabilities