Unlock Seamless Logical Replication: MySQL, Oracle & SQL Server to PostgreSQL

PostgreSQL has long excelled at native logical replication and is rapidly becoming a top choice for data analytics, thanks to its powerful features, flexibility, and extensibility. As organizations increasingly rely on diverse databases like MySQL, Oracle, SQL Server, and others, the need for seamless, high-performance replication into PostgreSQL—also known as Change Data Capture (CDC)—has grown significantly.

This session introduces SynchDB, an open-source project designed to enable high-performance logical replication from heterogeneous databases (MySQL, Oracle, SQL Server, and more) into PostgreSQL. Built on the Debezium Embedded Engine, SynchDB acts as a unified command center for managing data replication from multiple sources with different database types, eliminating the need for complex replication setups.
The talk will cover the technical challenges and design strategies behind making heterogeneous logical replication in PostgreSQL possible, including:

• Architecture decisions
• Bridging the gap between Java and C scopes
• Data types and expression transformations
• Error handling between Java and C
• Current status and next steps