The Iceberg Spec in Motion: v3 Features, v4 Proposals, and a Maturing Catalog Ecosystem

Apache Iceberg has become the default table format for the open data lakehouse, and the specification keeps moving fast. This talk walks through where things stand and where they head next.

We start with Iceberg v3 and what it actually changed for practitioners: new data types, deletion vectors, row lineage, and the features that reshape how engines read and write tables. From there we look ahead to the v4 proposals under active discussion on the dev list, what problems they aim to solve, and what they signal about the format’s direction.

The second half turns to the catalog layer, the part of the ecosystem that has grown most chaotic and most important. We map the current landscape, from the Iceberg REST Catalog spec to Apache Polaris and the wider field of managed and open options. You will leave understanding how the spec, the engines, and the catalogs fit together, and how to make architecture choices that hold up as v4 arrives.

Whether you run Iceberg in production today or plan your next platform, this session gives you a clear read on the format’s present and a grounded view of its future.