Materialize, SQL Database Startup, Raises Cash
Streaming data continues to enter the mainstream, its further adoption hindered primarily by the inability to run complex queries on multiple streams of data along with other engineering tradeoffs.
Startups seeking to make it easier to run those queries on streaming platforms without current tradeoffs are attracting financial backing as standard SQL frameworks emerge. Among them, streaming SQL database vendor Materialize announced a $32 million funding round this week led by Kleiner Perkins.
Thus far, New York-based Materialize has raised $40 million in two funding founds. Other investors include Lightspeed Venture Partners and the company founders’ former employers—Cockroach Labs and Datadog. The startup also includes veterans from DropBox, Pure Storage and YouTube.
The investments reflect the growing need to run complex queries on multiple data streams. The startup claims its approach offers millisecond response times while enabling customers to leverage in-house SQL skills.
The startup further claims the first standard SQL interface for streaming data aimed at engineers building complex queries that don’t require underlying micro-services infrastructure or specialized data skills. Materialize said its approach computes and incrementally maintains data as it is generated, making query results accessible as needed. “This approach provides businesses with correct answers in milliseconds – without interference from late-arriving data,” the company said.
Frank McSherry, co-founder and chief scientist of Materialize, is a former Microsoft (NASDAQ: MSFT) researcher who helped launch its parallel dataflow project dubbed Naiad. McSherry is credited with creating a computational model called “Timely Dataflow” that along with another model called “Differential Dataflow” serve as the basis of the Materialize platform.
The startup’s latest release earlier in November includes upgrades designed for production deployments and connections to other systems as well as improved Postgres compatibility and beta versions of source caching and tables.
Source caching is designed to reduce requirements for re-ingesting data while tables are used to load static data into the Materialize platform. The platform manages data in a table rather than relying on Apache Kafka or a file system.
Materialize said early customers are using its platform for applications such as real-time data visualization, financial modeling and advancing enterprise software services applications.
“Every business is going to be a real-time business within the next few years, but current approaches require compromises between cost, speed and skills to get there,” said Bucky Moore, partner at lead investor Kleiner Perkins.