Introducing Cloudflare Basin: an open, serverless data platform, now generally available

Cloudflare has rebranded its Cloudflare Data Platform as Basin, now generally available after a year in open beta. The platform is designed as a serverless data analytics solution, leveraging Apache Iceberg for open table format compatibility and R2 Object Storage for cost-efficient data access. Basin consolidates ingestion, storage, and querying into a unified system, enabling developers to collect data from multiple sources and analyze it without managing infrastructure.

Basin’s architecture prioritizes speed, openness, and cost efficiency, addressing challenges in modern data applications. Developers can create a Basin Catalog, set up Pipelines, and query data using Basin SQL within seconds. The platform scales by compacting metadata and data files to reduce I/O and distributing queries across Cloudflare’s global network, ensuring consistent performance even as datasets grow.

The platform supports Apache Iceberg-compatible engines like PyIceberg, DuckDB, Snowflake, and Apache Spark, allowing data portability and free egress across regions and cloud providers. Basin’s serverless model charges users only for data ingestion, processing, or querying, with no hourly or infrastructure costs. Early adopters, including Cloudflare’s billing and infrastructure teams, have used Basin for real-time e-commerce optimization, billing metrics reporting, and infrastructure telemetry analysis.

Basin consists of three core products: Pipelines for data ingestion, Catalog for storage and management, and SQL for querying. Pipelines can process up to 3GB/s per stream and transform data during ingestion to reduce storage footprint. Basin Catalog provides a fully managed Apache Iceberg REST catalog with automatic maintenance, while Basin SQL supports hundreds of functions for querying large datasets without provisioning clusters.