Build transactional data lakes with Apache Iceberg on AWS (hebrew)

I had a privilege to speak at AWS Summit 2024 in Tel Aviv.
My talk was about Apache Iceberg - an open table format which already proven to be a synonym for the modern data lake.

I've been working with Iceberg for several years and I am noticing a shift in AWS customers' interest in this technology.

In a past, customers were interested to learn how Apache Iceberg works and it’s core features. Customers evaluated if Iceberg is the right tool for them to use. Now they are looking for more practical knowledge on how to run Iceberg at scale. This is what we wanted to address in our talk for Summit.

We opened a session with Apache Iceberg overview, common challenges it solves and deep dive on how Iceberg works. Then he elaborated on how to proceed with migrating data to Iceberg using one of two existing migration strategies.

My part was to provide practical technics on how to optimize Iceberg ingestion and consumption, highlighting the tradeoffs and talking about maintenance operations and monitoring.

Finally we shared a fascinating story about Cloudinary data migration to Iceberg and what advanced technics were used by the data team to ensure Iceberg providing data warehouse performance while significantly lowering the cost.