Starburst Data, Inc. addresses a recurring challenge for large organizations: making use of data scattered across operational databases, analytical warehouses and cloud storage. Based in Boston, this American company offers an access and analytics layer designed to limit data movement and the proliferation of copies. Its website, starburst.io, presents its software, services and resources for using them.
Roots in distributed SQL
Founded in 2017 by Justin Borgman and others, Starburst is rooted in the history of Presto, a query engine initially developed at Facebook to analyze large volumes of data. A branch of this project, known as PrestoSQL, was later renamed Trino. This open-source engine forms the technological foundation of Starburst’s offerings.
The company has built its business model around making this technology production-ready: simplifying operations, improving performance and providing the security and support capabilities expected by enterprises. This trajectory illustrates a common model in software infrastructure, in which a company sells complementary services and features built around a community project.
A platform bridging federation and lakehouse architectures
The central principle is to execute SQL queries across different sources using connectors. A team can thus combine information from multiple environments without first having to transfer everything into a single warehouse. However, this federation approach eliminates neither the need for data preparation nor the performance constraints inherent in the systems being queried.
Starburst Enterprise is aimed at organizations that want to deploy and administer the platform in their own environment. Starburst Galaxy offers a managed cloud service experience to reduce the burden of operating the engine. The offerings include tools for access management, query monitoring and optimization, with capabilities that vary by product and configuration.
Starburst also operates in so-called “lakehouse” architectures, which bring together the flexibility of a data lake and the analytical uses of a warehouse. Support for open table formats such as Apache Iceberg makes it possible to query data stored in object storage systems. Target use cases include decision-making, analytical exploration and supplying data to applications that require SQL access to it.
What comes next?
What comes next will depend in particular on Starburst’s ability to reconcile technical openness, ease of operation and cost control. In competition with cloud warehouses and other lakehouse platforms, the company advocates an architecture in which storage and compute remain decoupled, with a strong emphasis on open standards.
The development of artificial intelligence applications could increase interest in governed access to heterogeneous data. For Starburst, the challenge will be to translate this demand into operational use cases without presenting federation as a universal solution: data quality, latency and access rights management remain essential conditions for success.