Building High-Performance Iceberg Data Platforms, Anywhere

​Description

​As data scales, cost, control, and performance become the primary concerns for data teams building out open table lakehouses. This talk walks through how to build a high-performance Apache Iceberg platform that gives you all three, using open standards as the foundation so you can run wherever your business needs it.

What the session covers

• ​What cloud-native design actually means as a principle: fast, easy scaling, not dependence on any specific cloud.

• ​How Iceberg's optimistic concurrency model works under the hood, and why atomic commits matter when multiple engines write to the same table.
   How Iceberg gives object storage table-level guarantees, schema tracking, snapshot isolation, and safe concurrent writes, without locking data into one engine.

• ​What an Iceberg catalog actually does, and the options and tradeoffs for how to deploy it in a lakehouse.

​What the audience will take away

• ​A clearer mental model of how catalog, metadata, and storage relate to each other in a lakehouse.

• ​Considerations for building a lakehouse that supports your workloads from BI today to AI tomorrow and any workloads you add in the future.

• ​Questions to bring back to their own team about Iceberg architecture decisions.
  
  Bring questions. The last stretch is open Q&A.