metamorworks - stock.adobe.com

DataStax acquires ML vendor as part of AI evolution

The acquisition will let customers apply machine learning to real-time events. It is a natural step toward AI technology for the vendor as it looks to stay competitive.

After acquiring machine learning vendor Kaskada, database and real-time event streaming vendor DataStax is shifting its focus to AI and ML.

The vendor, based in Santa Clara, Calif., revealed the acquisition on Jan. 12 and said it plans make the Kaskada technology open source first and then offer a new ML cloud service later this year.

In somewhat of a rebranding, DataStax is now calling itself a "real-time AI company." Terms of the acquisition were not disclosed. Kaskada had raised $9.8 million since its founding in 2018.

Kaskada's technology processes large amounts of data while updating features for ML models over time.

Founded in 2010, DataStax is known for its Astra DB database-as-a-service based on Apache Cassandra.

DataStax's acquisition of Kaskada is its fifth since its DataStax was founded in 2010 and its first since acquiring cloud messaging vendor Kesque in January 2021.

Adding Kaskada to DataStax's portfolio will provide organizations an environment to develop applications inspired by AI, according to DataStax.

A natural step

The acquisition is a natural step for the data vendor, according to Carl Olofson, an analyst at IDC.

Since Cassandra is an SQL database, it has considerable flexibility, including the ability to set up training data for ML, Olofson said. However, many users who are not ML experts may need help with using it to set up training data for ML.

This is where Kaskada fits in. The Seattle-based vendor provides users with tools to train behavioral ML models directly from event-based data produced from the tracking and analysis of the interactions between customers and a product.

This follows a trend that a number of other database companies have already been following, which is building machine learning tools, autoML and those kinds of things into the database.
Carl OlofsonAnalyst, IDC

"Having a tool that sets up [the training data] for you is of great value," Olofson said. "This follows a trend that a number of other database companies have already been following, which is building machine learning tools, autoML and those kinds of things into the database."

This acquisition also lets DataStax keep competing with the likes of Microsoft Azure Cosmos DB and databases from AWS and Oracle.

"It's becoming fairly common to offer some level of machine learning capability combined with a database," Olofson said. He added that DataStax is increasingly also looking to compete on the analytics side -- a fast-growing aspect part of the database market.

"This puts them in the game for a new and relatively new and evolving market," he continued.

A possible challenge for DataStax is how it will incorporate the AI and ML technologies.

"Whenever you're introducing another format, that's a challenge to integrate. That's always been a challenge and will continue to be a challenge," Olofson said.

Dig Deeper on AI infrastructure