Improved Performance with Starburst ODBC Driver Version 3

Starburst Enterprise and Starburst Galaxy already deliver success for critical workloads in AI, analytics, and data applications. For these use cases to function, data access is key. In fact, data access is a central cornerstone of Starburst’s value. 

Optionality: The core Starburst value

Starburst provides organizations with true choice over their data, allowing for as much data centralization or decentralization as needed. Whether powering data analytics, AI workloads, data applications, or other use cases, data access is the key to Starburst’s success and business impact. 

This means that the drivers and connectors that facilitate data access are the central building blocks of Starburst itself. In this article, we will explore the ODBC driver. We’ll unpack how data is accessed using ODBC, why it matters, and how Starburst is improving access with the new version 3 driver.

 

What is ODBC? 

Desktop applications utilize various APIs to interact with data, including JDBC and ODBC. In the domain of Windows BI tools, ODBC is the ubiquitous API for many, including Microsoft Power BI and Tableau.

The Starburst ODBC driver enables enterprises to track and optimize their business using their established tools and processes.

How to access Starburst with ODBC?

Your Starburst cluster’s access to data is available via ODBC when you use an application that supports ODBC. Typically, either you or an administrator will need to install the ODBC driver and then specify the Starburst cluster as a data source. From there, users can query the data using SQL or build data applications or AI workflows.

ODBC is widely used on the Windows operating platform, but it is also used on macOS and Linux.

 

How we enhanced Starburst’s connection to ODBC data sources

ODBC Version 3 marks a new beginning for client connectivity to Starburst. It is a complete redesign from ODBC version 2.

For this release, Starburst moved from repackaging a third-party driver to building the driver with our internal Engineering team. The team’s proximity to and knowledge of Starburst enabled faster innovation while also delivering a more performant and versatile driver. Most notably, the new driver leverages the new improvements provided by the Trino spooling protocol. This helps drive major improvements in driver performance. 

How does the Starburst ODBC driver work? 

The diagram below illustrates the operation of an ODBC driver. An ODBC application uses either a connection string (e.g., Power BI) or a data source name (e.g., Tableau) with the Driver Manager to connect to access the full range of data available from Starburst.

For more information, see the ODBC driver documentation page. It outlines the differences between version 2 and version 3.

What makes ODBC version 3 important? 

ODBC version 3 was built as a drop-in replacement for version 2. No change is required to your environment beyond installing the new driver. Internal benchmarks indicate that version 3 is as performant as version 2 when used with Trino’s original direct protocol.

When using Starburst Enterprise 468-e LTS onward, or with Starburst Galaxy, the new ODBC driver leverages the speed enhancements of the spooling protocol. Internal benchmarks showed improvements of up to 400% on the transfer of result sets over version 2. 

For example, an extract of 200 columns and 1.7 million rows (totalling approximately 1.7 GB) from a financial services database, using a popular BI tool, demonstrated a 3.4x improvement in transfer time with no other changes. 

The chart below compares the query time between the v2 and v3 drivers. Notice the substantial performance improvement. This change is a result of the spooling
protocol’s use on the object store instead of the coordinator to return results.

 

Compatibility of Starburst’s new ODBC driver 

Which versions of Starburst support the new ODBC driver? 

While Starburst Enterprise 468-e LTS or Starburst Galaxy are required to benefit from maximum performance gains due to the spooling protocol, the new version 3 ODBC driver is backwardly compatible with supported versions of Starburst Enterprise. This means that all current Starburst customers will benefit, and those using newer versions of Starburst Enterprise and all versions of Starburst Galaxy will also benefit from enhancements due to spooling. 

ODBC driver release schedule. 

The version 3 driver was in public preview in the first half of 2025. As with all changes to drivers, we encourage organizations to test the new release in their integration environments in preparation for production. 

Presently, version 3 is only supported in Windows. Support for Linux and Mac is planned for a future update. Beyond this, version 2 is still supported and compatible across all operating systems.

For more information on support for the ODBC version 3 driver, please contact your Starburst sales team. 

 

Why ODBC data matters today

Starburst has made significant investments in the next generation of our ODBC driver, ensuring its resilience while also improving performance by leveraging spooling protocol enhancements within the query engine itself. 

Starburst’s investment in ODBC has a direct benefit for our customers. Try it for yourself to see how these improvements will aid in your GenAI, BI, and data applications today.