Iceberg REST Catalogs

"Iceberg REST Catalogs: Managing Table Metadata Across Engines and Clouds"
Iceberg has made analytics tables portable, but in real platforms the hardest problems live above the files: coordinating metadata safely across many writers, many query engines, and many clouds. This book targets experienced data engineers, platform architects, and lakehouse operators who need a catalog layer that behaves like a reliable control plane—not a fragile integration point—while keeping interoperability and correctness front and center.
You’ll build a metadata-first mental model of Iceberg tables (manifests, metadata.json, snapshots, and atomic commits), then move into the catalog’s real contract: identity, namespace strategy, commit coordination, and the boundaries between what the catalog must guarantee and what engines must enforce. A protocol deep dive covers REST lifecycle operations, optimistic concurrency, idempotent retries, compatibility and graceful degradation, plus an error taxonomy that maps HTTP failures to root causes. From there, the book tackles multi-writer correctness, retention and cleanup, incident-grade failure modes, and upgrade choreography when fleets can’t move in lockstep.
Security and operations are treated as first-class: AuthN/AuthZ models that work across engines, vended credentials for object storage, auditing for compliance, and SLO-driven observability. Practical chapters show Spark and Trino integration patterns, then guide you through selecting and deploying REST catalog services and executing a safe migration playbook—including dual-run, cutover, rol