Starburst

Starburst is an enterprise analytics platform built on Trino (formerly known as PrestoSQL), the fast, distributed SQL query engine originally developed at Facebook. By providing a commercially supported, enterprise-hardened version of Trino with additional governance, security, performance, and management features, Starburst enables organizations to query data across heterogeneous data sources without moving or copying data into a central repository.

The data mesh and data lakehouse architectural patterns that Starburst supports are based on a fundamental principle: data should be queried where it lives rather than copied to a central data warehouse. Starburst implements this through its catalog system, which provides connectors to dozens of different data sources including Amazon S3, Google Cloud Storage, Azure Data Lake, Apache Hive, Apache Iceberg, Delta Lake, Apache Hudi, as well as traditional databases like PostgreSQL, MySQL, Oracle, SQL Server, MongoDB, Elasticsearch, Kafka, and many others. Each connected data source appears as a separate catalog in the SQL interface.

Cross-source queries are one of Starburst's most powerful capabilities. Analysts can write a single SQL query that joins tables from different catalogs, such as joining a PostgreSQL operational database table with a Parquet file in S3 with a Redshift table, all in one query. The Trino query engine distributes the query execution, pushes down predicates to each source system where possible, and assembles the results, all transparently from the user's perspective.

The Starburst Data Products feature introduces a governed data access layer built on top of the raw Trino query engine. Data products are defined views with descriptions, owners, access policies, and quality metrics that present curated datasets to data consumers. This catalog-based discovery and governance capability helps organizations build well-organized, trustworthy data ecosystems at scale.

Column masking and row-level security policies in Starburst provide dynamic data access control. Column masking applies obfuscation rules to sensitive columns such as credit card numbers and email addresses based on the requesting user's role. Row-level security restricts which rows each user can see based on policy expressions. These controls are enforced at query time, ensuring that access policies are applied consistently regardless of which BI tool or query client is used.

Starburst Galaxy is the fully managed cloud service that provides Starburst on AWS, Azure, and Google Cloud with automatic scaling, serverless query execution, and simplified cluster management. Starburst Enterprise is the self-managed version for deployment on Kubernetes in customer-controlled infrastructure.

Starburst integrates natively with popular BI tools including Tableau, Looker, Power BI, Superset, and others through JDBC and ODBC drivers, enabling existing BI workflows to leverage Starburst's cross-source query capabilities without changing the BI tool.