March 2026 Summaries

Apache Spark, a leading open data processing framework utilized by over 20,000 companies, has seen Kubernetes emerge as a favored platform for running open-source Spark, second only to Databricks. While managing Spark on Kubernetes offers cost savings and infrastructure standardization, it often requires significant effort and entails performance trade-offs compared to managed platforms. The introduction of the Quanton Kubernetes Operator aims to bridge this gap by providing a solution that enhances Spark job performance by 3-4 times without necessitating code changes, offering organizations the flexibility and control of self-managed infrastructure alongside the performance benefits typically associated with managed services. Quanton, designed specifically for lakehouse ETL and Spark workloads, integrates seamlessly into existing Kubernetes setups and offers a comprehensive Spark experience through the Onehouse UI, providing both performance improvement and cost efficiency.

OneSync's new feature, cross-catalog permission translation, enhances its capability by allowing automatic translation and enforcement of access control policies across multiple platforms such as Microsoft OneLake, AWS Lake Formation, Databricks Unity Catalog, and Snowflake. This innovation addresses the challenge of maintaining consistent data access permissions across a multi-cloud and multi-platform environment, where different teams might use different data services. Previously, users had to manually duplicate permission models across different catalogs, which was inefficient and posed compliance risks. With OneSync, permissions can now be defined in one place and automatically translated and enforced across all connected catalogs, ensuring consistent access control. The system supports both database-level and table-level permissions and is set to include column-level and row-level permissions in the future, enhancing data governance by allowing more granular control, such as column masking and regional data restrictions. This development is crucial for organizations aiming for an open data architecture, as it ensures that governance models are preserved with data across various platforms, maintaining security and compliance without needing to consolidate onto a single platform.

Onehouse Cloud has expanded its availability to Microsoft Azure, complementing its existing support for Amazon Web Services and Google Cloud, thereby providing a comprehensive solution across the three largest public cloud providers. This expansion allows Azure users to leverage Onehouse's cost-effective and efficient data lakehouse infrastructure, which integrates deeply with Azure services such as Azure Data Lake Storage, Azure Event Hubs, and Azure Database for PostgreSQL. The platform offers features like OneFlow for data ingestion, SQL and Spark jobs powered by the Quanton engine, and table optimization through LakeBase, providing robust solutions for real-time analytics and AI workloads. Onehouse's partnership with Microsoft has resulted in the co-creation of Apache XTable, enabling interoperability among various table formats and enhancing the Onehouse offering on Azure. The service is deployed within an Azure VNet, ensuring data security and control while allowing seamless integration with existing Microsoft OneLake users and facilitating an easy setup for new lakehouse users.