Today, we are excited to announce Starburst Enterprise 479-e.1 LTS. This release advances our commitment to supporting the foundations of AI, extends our ongoing leadership on Iceberg, and makes it easier than ever to manage your data sources through enhancements to SQL-based orchestration and engine performance.
Collectively, 479-e introduces significant improvements in the following areas:
- Starburst Icehouse improvements with enhanced manual and automatic refreshes for Iceberg Materialized Views.
- Engine performance breakthroughs, yielding up to 2x faster performance over open-source Trino.
- Expanded connectivity leveraging the Great Lakes connector and the new OpenAPI connector.
- Data Product enhancements, such as support for sharing data products across clusters and cron-based refreshing, as well as Data Products as Code (now in Private Preview).
- Modernized Infrastructure with full support for Java 25.0.2 and enhanced Starburst Warp Speed caching.
Starburst advances the foundations of AI
Starburst continues to deepen its status as the ideal foundation for AI workloads by focusing on the performance and orchestration layers that sit beneath agentic workflows. In this release, we have introduced three core enhancements involving Common Table Expression (CTE) reuse, SQL Jobs, and connectivity.
CTE reuse for complex AI reasoning
Common Table Expression (CTE) reuse directly supports AI workloads, allowing the engine to identify repeated subqueries within a complex query, execute them once, and reuse the results. This significantly reduces CPU and I/O load.
For AI agents performing high-context retrieval or traversing complex schemas, this results in faster query times and more efficient resource utilization.
Multi-Statement SQL Jobs now include built-in orchestration for AI context
We have introduced Multi-Statement SQL Jobs, which let you execute up to 200 SQL statements consecutively through the Starburst Enterprise UI or REST API.
Instead of triggering individual queries manually, users can now orchestrate a full logic flow within one automated, scheduled job. This provides a robust, built-in layer for preparing the governed data sets that fuel AI models and business applications.
Connectivity for the Great Lakes Connector and APIs
New to Starburst Enterprise, the Great Lakes connector provides a unified path for accessing data across various object storage formats. Additionally, the new OpenAPI connector allows users to query data gated behind REST API endpoints using standard SQL.
This expands the reachable data for your AI agents, allowing them to reason over data that previously required custom code or disjointed architectures.
Starburst makes Iceberg table maintenance easier than ever
Starburst continues its ongoing commitment to Apache Iceberg as the foundation of our Icehouse architecture. We believe Iceberg represents the best foundation for AI, and 479-e doubles down on making that foundation production-ready in a number of different ways.
Materialized View automatic refresh
In the 479-e LTS, manual and automatic refreshes for Iceberg Materialized Views are now generally available. This allows organizations to offload complex joins onto optimized, reusable views.
With the added support for cron expressions, you can now schedule these refreshes with granular precision, ensuring your AI agents are always working with the most current data.
Performance and interoperability
We have implemented engine-level optimizations to reduce potential memory pressure when querying Iceberg tables with highly nested fields.
Furthermore, 479-e improves the compression logic to ensure full interoperability between Trino and Spark, keeping your data truly open and accessible across your entire toolchain.
Enterprise-grade management and governance
This release also includes enhancements to how Data Products are managed and how clusters interact within a global data architecture.
Sharing Data Products across clusters
In most organizations, data isn’t confined to a single cluster. To further support this, we have added public preview support for sharing data products across clusters, allowing teams to consume governed, high-quality data assets regardless of their location.
This ensures that a curated, business context layer data product remains the single source of truth for every department.
Modernized Infrastructure with Java 25
All Starburst Enterprise versions have been upgraded to run on Java 25.0.2. This ensures that our customers benefit from the latest performance improvements, security patches, and long-term support provided by the Java ecosystem.
Critical breaking changes for 479-e LTS
This release includes several critical changes that require attention during the upgrade process. Please consult the full documentation for the complete list of breaking changes.
Starburst Warp Speed changes
Warp Speed now requires local SSD storage on the coordinator in addition to worker nodes. Furthermore, several features, including Fast Warmup and Index Resiliency, have been removed.
Note: Users must remove the associated configuration properties or their cluster will fail to start.
Mandatory security masking for custom connectors
As of this release, the security masking implementation for custom connectors is no longer optional. In previous releases, the default implementation of the ConnectorFactory #getSecuritySensitivePropertyName method was abstract.
Now, Custom plugins must provide their own implementation of the property-masking method to identify security-sensitive information. This change ensures the engine can properly mask sensitive values, but it requires an update for any customer who maintains their own custom plugins to avoid compilation or runtime issues.
Apache Ranger support changing
As of 479-e, integration with Apache Ranger Admin 2.0.0 is no longer supported. Users must upgrade their Ranger deployment to version 2.1.0 or higher.
Unity Catalog property renames
Several Delta Lake Unity Catalog managed table configuration properties have been renamed. For example, hive.metastore.unity.catalog-owned-table-enabled is now hive.metastore.unity.catalog-managed-table-enabled.
Upgrade to Starburst Enterprise 479-e LTS Today
The 479-e LTS release marks a significant milestone in enabling organizations to automate Iceberg, orchestrate complex SQL workloads, and scale their AI foundations.
View the full release notes here.
