Download the paper: Benchmarking Native In-Database TPCx-AI at 100 Terabytes in Ocient Hyperscale Data Warehouse: An Eight-Use-Case Study of Classical and Statistical ML on Relational Primitives
As enterprise data volumes grow, traditional machine learning workflows often require moving large datasets into external processing frameworks, adding transfer time, infrastructure overhead, and governance complexity. This paper presents Ocient’s approach to executing classical and statistical machine learning directly in the database, using standard SQL and native relational primitives.
Published in the Proceedings of the VLDB Endowment, the paper details how Ocient executed eight TPCx-AI use cases at a 100 TB scale factor, with performance evaluated across six data scales from 1 GB to 100 TB. It also examines the underlying architecture and compares results with algorithm-matched Apache Spark implementations.