Introducing the Great Lakes Connector for Starburst Enterprise

Starburst, the data foundation that powers analytics and AI, now supports the Great Lakes connector in Starburst Enterprise, a unified connector for all your data lakehouse table formats.

Organizations rarely live with a single table format. If you’re migrating from Hive to Apache Iceberg, introducing Delta Lake for specific workloads, or supporting multiple engines side by side, you quickly end up with multiple catalogs, duplicated configuration, and confused end users. In this context, the modernization of the lakehouse becomes a long-running migration project spanning multiple years, teams, and use cases.

Enter the Great Lakes Connector provides

With the Great Lakes connector on Starburst Enterprise, you can finally simplify that reality. One connector, one catalog, and a consistent SQL experience across Hive, Iceberg, Delta Lake, and Hudi. 

Check out this demo below, showing the connector in action. 

This capability builds on our proven Great Lakes connectivity in Starburst Galaxy and goes beyond the basic Lakehouse connector available in open‑source Trino — giving you a production‑ready and feature‑rich way to run multi‑format data lakes at scale.

Unpacking the challenge of many table formats and too many catalogs

Let’s start by exploring the problem that the Great Lakes Connector solves. Most large enterprises fit into one of these patterns:

• You’re migrating from Hive to an open table format such as Iceberg or Delta Lake, often on a use-case-by-use-case or team‑by‑team basis.
• You’re adopting Iceberg or Delta Lake for certain workloads while keeping existing Hive tables.
• You expect this transition to last years, not months.

Today, those realities usually mean three predictable outcomes. 

The result is a higher total cost of ownership and slower migrations. Teams spend time managing plumbing instead of delivering value.

How the Great Lakes connector for Starburst Enterprise solves this problem 

The Great Lakes connector is Starburst’s umbrella connector for open table formats. It combines the capabilities of our Hive, Iceberg, Delta Lake, and Hudi connectors into a single, unified connector, backed by a shared file system and metastore such as Hive Metastore or AWS Glue.

The Great Lakes Connector in action

With Great Lakes in Starburst Enterprise you can:

• Query and write across multiple table formats from one catalog.
• Create new tables by choosing a type, such as iceberg, hive, or delta, without changing catalogs.
• Keep using your existing object storage and metastore configuration — Great Lakes sits on top of what you already have.

A simple example illustrates the experience:

-- Create an Iceberg table in your Great Lakes catalog
CREATE TABLE great_lakes.customer (
    customer_id varchar,
    region      varchar
)
WITH (
    type = 'iceberg'
);

-- Later, just query it like any other table
SELECT * FROM great_lakes.customer;

The connector handles the format selection behind the scenes. Your users write SQL. The Great Lakes Connector picks the right table format and underlying connector for the job.

Built on proven Great Lakes connectivity in Starburst Galaxy

This isn’t a brand‑new idea. Great Lakes connectivity has been available in Starburst Galaxy, where it powers a simple yet powerful user experience:

• Connect once to your Amazon S3, Azure ADLS, or Google Cloud Storage (GCS) object storage.
• Use one connector to read and write Hive, Delta Lake, and Iceberg tables.
• Specify the desired table type in SQL, then query everything transparently.

End users don’t need to know which table format is behind a dataset. Instead, they just run SQL against a single catalog. Great Lakes in Starburst Enterprise brings that same simplicity and consistency to your self‑managed deployments, aligning the experience across Starburst Galaxy and Starburst Enterprise.

Great Lakes vs. Trino’s Lakehouse connector

Open‑source Trino includes a Lakehouse connector that aims to achieve a similar goal–a single connector for multiple table formats. However, there are important differences in capability, maturity, and operational readiness.

We conducted an internal evaluation and came to the following conclusions. 

Trino Lakehouse connector

The standard Trino Lakehouse connector handles basic read and write operations, but lacks several advanced integration features. Specifically, it does not support system tables, procedures, or table procedures, nor does it allow the use of scalar or table functions.

Great Lakes connector in Starburst

The Starburst Great Lakes connector builds on Starburst Galaxy’s high-performance architecture to provide enhanced operational and metadata capabilities. It accelerates data access through parallel streaming of relation metadata, relation type caching, and deep integration with our warp speed planning layers. 

Beyond raw performance, the connector supports advanced features like the UNLOAD table function and location-aware access control, ensuring a consistent, enterprise-grade experience as these capabilities expand across the platform.

Overall results of Starburst Great Lakes Connector vs Trino Lakehouse Connector

In short, the Trino Lakehouse connector is a good starting point in OSS. However, it is limited and can face issues when implemented in enterprise production environments without sufficient support. 

In contrast, the Great Lakes Connector is designed to be enterprise-ready out of the box. It is therefore the implementation we recommend and support in Starburst Enterprise.

How to get started with the Great Lakes Connector

Wondering how you can get started with the Starburst Enterprise Great Lakes Connector? If you’re an existing Starburst Enterprise customer:

1. Upgrade to the Starburst 479-e.1 Feb LTS release that includes the Great Lakes connector.
2. Configure a Great Lakes catalog using your existing object storage and metastore settings.

-- Create a Great Lakes catalog using dynamic catalog
CREATE CATALOG great_lakes_s3
USING great_lakes
WITH (
"hive.metastore"='glue'
"fs.native-s3.enabled"='true'
);

     3. Start creating new tables with type=’iceberg’, type=’delta’, or type=’hive’ and query them from a single catalog, just as your Galaxy peers do today.

The Great Lakes connector in Starburst Enterprise makes multi‑format data lakes feel simple. This lets you focus on delivering value from your data, not fighting your catalogs.