Skip to content
Lakehouse Catalogs & Governance Last updated: May 29, 2026

REST Catalog Credential Vending

A security architecture where the Iceberg REST catalog dynamically issues short-lived, scoped storage credentials to query engines, eliminating static bucket credentials.

credential vendingtoken vendingrest catalog security

REST Catalog Credential Vending

REST Catalog Credential Vending is a security pattern defined in the Apache Iceberg REST Catalog specification. In traditional data lakes, every query engine (such as Spark, Trino, or Dremio) requires long-lived read and write access credentials configured locally to read files from object storage. With credential vending, the REST catalog server acts as the single source of security truth. It generates temporary, tightly scoped credentials and sends them to the query engines at runtime.

How Credential Vending Works

When an engine wants to read or write to an Iceberg table, it follows a multi-step communication flow:

  1. Table Request: The query engine sends a request to the REST catalog to load a table (e.g. /v1/namespaces/{ns}/tables/{table}).
  2. Access Verification: The catalog server verifies that the client has the appropriate permissions to access that table.
  3. Credential Generation: If authorized, the catalog requests temporary credentials from the underlying cloud provider (such as AWS STS for S3, GCP IAM for GCS, or Azure Entra ID for ADLS). These credentials are restricted to the exact bucket path where the table data and metadata reside.
  4. Metadata and Credential Response: The catalog returns the table metadata location along with the temporary credentials in the HTTP response.
  5. Direct Storage Scan: The query engine uses the temporary credentials to access the object storage directly, completely bypassing the catalog server for the heavy lifting of data scanning.

Security and Architectural Benefits

This approach addresses several common security challenges in lakehouse environments:

๐Ÿ“š Go Deeper on Apache Iceberg

Alex Merced has authored three hands-on books covering Apache Iceberg, the Agentic Lakehouse, and modern data architecture. Pick up a copy to master the full ecosystem.

โ† Back to Iceberg Knowledge Base