Starburst Debuts ‘Icehouse’ Managed Data Lake Service

Starburst is touting the new lakehouse platform, with its support for the Apache Iceberg table format, as a high-performance, more cost-effective alternative to competing cloud data lake and data warehouse services for data management, analytics and AI tasks.

Starburst is launching a fully managed “Icehouse” data analytics service, based on its Galaxy cloud data lakehouse platform, that combines the Trino distributed SQL query engine with the Apache Iceberg table format for analytic datasets.

This week also marked the availability of the Dell Data Lakehouse, a previously announced collaboration between Starburst and Dell Technologies through which Dell has combined its PowerEdge servers with the Starburst data analytics engine software to create a turnkey data lakehouse system.

The developments mark Starburst’s continued expansion of its reach in the data warehouse/data lakehouse arena – Wednesday the company said that Starburst Galaxy has recorded 3x year-over-year growth in both active customers and usage volume.

The news also comes a week after Boston-based Starburst, founded in 2017, expanded its C-suite with the hire of several new executives – including new President Steven Chung – as the company lays the groundwork for a future initial public offering.

“This is about getting us ready for an IPO, to be perfectly honest,” co-founder and CEO Justin Borgman (pictured) said in an interview with CRN. “Every startup goes through its journeys [and] certain executives are right for certain periods of that journey. As you transition to a later-stage company and prepare to go public, it’s a different business.”

Starburst last raised funding in a $250 million Series D round in February 2022 that brought its total funding to $414 million and its then-valuation to $3.35 billion.

The debut of Galaxy Icehouse at this week’s Data Universe 2024 event in New York City is Starburst’s latest expansion of its data analytics offerings. Galaxy Icehouse combines the open-source Trino SQL query engine – the core of the Starburst platform – with the Apache Iceberg data table format to provide a fully managed data lake service that the company says provides “powerful scalability, cost-effectiveness and query performance” without the burden and cost of building and maintaining a custom system and “without the risk of vendor lock-in.”

Borgman emphasized the cost-performance benefits of Galaxy Icehouse compared to traditional cloud-based data warehouses such as Amazon Web Services’ Redshift. He also cited its effectiveness in handling compute-intensive data transformation and preparation workloads.

Galaxy Icehouse supports near real-time, petabyte-scale data ingestion into Iceberg managed tables, according to Starburst. Data and development teams can use SQL to prepare and optimize data for production analytics, the company said, along with the auto-tuning capabilities in Starburst Warp Speed to improve query performance.

Data table formats are critical for organizing and managing large datasets for analytical tasks and are a key component of data lake platforms. The three most popular – and competing – formats are Apache Iceberg, Apache Hudi and Delta Lake – the latter developed by data lakehouse giant Databricks. (Iceberg was originally developed by Netflix to work with the Apache Hive Hadoop project.)

Borgman said Starburst supports all three formats, “but we think Iceberg is winning this battle” and will become the industry standard. “And we think that because we see a broad ecosystem of vendors choosing to support Iceberg at this point,” he added, pointing to data cloud giant Snowflake as an example.

Starburst works with global and regional systems integrator partners (along with OEM and technology partners) and Toni Adams, Starburst vice president of partners and alliances, said in the interview with CRN that more than 50 percent of customer engagements for new, renewal and expansion sales involve a partner.

Adams said partners will benefit from the foundation that Starburst is creating with Galaxy Icehouse for data interoperability and managing and querying data, providing partners with “a competitive advantage when they are building solutions around data,” especially in the fast-growing AI area.

“Today, it's all about AI and AI management, meaning AI and data and data management are intertwined,” Adams said. “There is no AI without data. So all of our partners are looking to how they meet customer needs around ML and AI and how that data is managed. And that is best managed through the Icehouse format.”

Last week Starburst said it had hired Stephen Chung as Starburst’s first president. He was previously president at Delphix, a DevOps data management tech developer that is being acquired by Perforce.

He previously was senior vice president of worldwide sales, services and business development at DevOps company PagerDuty, which went public in 2019, and before that was senior vice president of worldwide sales and business development at Demandware, which Salesforce bought in 2016. He has also held management jobs at Symantec, Microstrategy and PwC Consulting.

Starburst also hired Tobias Ternstrom as chief product officer, joining from Nutanix where he was vice president of platform services. And Adam Ferrari was hired as senior vice president of engineering, coming to Starburst from product information management company Salsify where he held the same position.