On November 4, 2026, Oracle released version 23.26.4 of GoldenGate for DAA, the Distributed Applications and Analytics edition, to all customers. The release can read changes from Snowflake, and it works with Oracle AI Data Catalog to manage Apache Iceberg tables.
It also certifies newer Confluent and Apache Kafka versions. Together, these changes widen the set of systems that GoldenGate can read from and write to, which matters for teams that move data between operational stores, streaming platforms and lakehouses.
What Is New in 23.26.4
The release has two headline additions and a set of new certifications. The additions are a Snowflake source and a native catalog integration for Iceberg tables. The certifications cover Confluent and Kafka.
- Snowflake Capture for initial loads and real-time replication.
- Native Iceberg catalog support through Oracle AI Data Catalog.
- New Confluent and Kafka certifications.
Snowflake Capture
With this version, Snowflake tables can act as a source of change data, which GoldenGate then delivers to its supported targets. Until now, Snowflake was usually where data ended up, not where changes began.
Oracle says both the first bulk load and the continuous replication that follows are supported. One tool can therefore seed a target and then keep it current. Setup details sit in the Snowflake Capture documentation.
Oracle AI Data Catalog for Iceberg
GoldenGate now integrates directly with Oracle AI Data Catalog for Iceberg. Per Oracle, the built-in integration supports lakehouses in Iceberg format that update in real time, stored on OCI Object Storage, Azure Storage or Oracle ZFS.
In practice, a catalog tells query engines where the tables and their metadata live. Pointing GoldenGate at the catalog means replicated data can arrive as managed Iceberg tables. Our earlier story on Oracle AI Lakehouse Iceberg catalog federation covers the lakehouse side of the same direction.
New Certifications
Oracle tested Confluent 8.2 and 8.3 for use on both the source and target side. Apache Kafka 4.3 passed the same test in both roles, running with the 4.x dependency set.
Certification means Oracle has tested these combinations, so you can upgrade Kafka or Confluent with less worry about breaking a replication pipeline. The Oracle GoldenGate certification matrix lists the full details.
What It Means for Data Engineers
If your company uses Snowflake for analytics and Oracle Database or Kafka elsewhere, this release removes one reason to write custom extract jobs. You can treat Snowflake changes as a stream and send them on.
The Iceberg catalog support is useful if you are building an open table format lakehouse and want Oracle's catalog to manage it. Test carefully, since the source gives feature names but no performance numbers or limits. Check the documentation for supported data types and target settings before planning production use.
How to Get the Release
Download is through Oracle Support and needs a current support contract. Oracle names two patch numbers for the release.
- 40171400 for the complete install of 23.26.4.
- 40171393 for the patch of 23.26.4.
Oracle also links release notes, 26ai documentation, the certification matrix and LiveLabs workshops for hands-on practice. Read the release notes first to confirm upgrade steps for your current version.
Conclusion
This release makes Snowflake a change source, connects to Oracle AI Data Catalog for Iceberg, and adds Confluent 8.2, 8.3 and Kafka 4.3 to the tested list. Review the release notes and certification matrix, then trial the new connectors in a test environment before you roll them out.
Sources
See my articles more often when you search Google.
