Starburst and Varada partner for Trino data platform effort

Two of the leading vendors in the open source Trino market are coming together in a new partnership between Starburst and Varada, made public March 16.

Trino is the rebranded name for what had been known as PrestoSQL until December 2020.

PrestoSQL was one of two variations of the open source Presto distributed SQL query engine originally created at Facebook. The other version of Presto is known as PrestoDB and is managed as an open source project at the Linux Foundation.

Both Starburst and Varada are founding members of the Trino project, though the two vendors have taken different approaches to the technology.

Starburst, based in Boston, has an enterprise platform that enables users to manage and query data. Starburst recently gained the attention of venture capital firms, including Andreessen Horowitz and Salesforce Ventures, both of which participated in  Starburst raising a $100 million Series C round on Jan. 6.

Meanwhile, Varada, based in Tel Aviv, Israel, raised a $12 million Series A in September 2020 and introduced its data platform in December 2020. Varada has focused on query acceleration and data virtualization with its Trino-based platform.

The new partnership will help to enable the vendors to work together to provide organizations with technology from both Varada and Starburst to optimize the management, delivery and query of data using Trino.

IDC analyst Stewart Bond said that though Varada and Starburst compete in some ways, the partnership focuses on where the two vendors can complement each other.

Bond noted that Starburst provides a data access layer to multiple types of disparate data sources that can include data lakes, data warehouses and operational data sources.

"Databases associated with analytical and operational workloads typically have native query acceleration capabilities that include indexing and optimized query plan execution, but data lakes do not," Bond said. "Varada provides query acceleration for data being managed in a lake, through indexing, caching and acceleration that adapts query execution based on parameters and analysis of workloads."

The partnership is a reflection of the demand for technologies that make data lakes more consumable.

While Starburst offers a high-performance query engine for a data lake, with the Varada partnership, query processing on top of the data lake will be faster and configurable, Bond added. In cases in which a Starburst customer has a mixed set of databases, and a data lake that must be queried as a federated dataset for analytics, Varada will be attractive for accelerating the performance of the data lake queries, he said.

"The partnership is a reflection of the demand for technologies that make data lakes more consumable," Bond said. "Varada is improving the performance of data lake queries for analytical purposes, and Starburst is enabling data lakes to be combined with analytical and operational databases in analytical workloads."

Why Starburst and Varada are partnering on Trino

Justin Borgman, chairman and CEO of Starburst, said he sees partnering with Varada as a way to help his own customers with query indexing and acceleration. As to why Starburst doesn't build its own technology that does what Varada provides, or even just acquire Varada outright, Borgman said the pact is really mostly focused on speeding go-to-market efforts.

Justin Borgman

"Partnering always makes sense, at least from my standpoint, as a way of getting to market quickly," Borgman said. "Building takes time and even acquisitions take time in terms of integration and so I think partners are always the fastest path to market."

Eran Vanounou

Eran Vanounou, CEO of Varada, echoed Borgman's view, saying he sees his platform's focus as complementary to Starburst's. Vanounou emphasized that Varada's concentration is not on cluster management of distributed data nodes, an area in which Starburst specializes.

"What we do really well is indexing and query acceleration," Vanounou said.