Apache Iceberg on AWS
Open at every layer of the analytics and AI stack
Bring every engine and agent to a single copy of governed data
When data is spread across siloed systems and locked behind proprietary platforms, every new tool means another copy of data, another integration, and another governance gap. Data teams spend hours on preparation while the business waits to move forward. And as AI agents add new access patterns, the problem only grows.
Leading data teams are building open data architectures with Apache Iceberg on AWS to bring every tool and agent to a single, governed copy of data. AWS offers the broadest native Iceberg support of any major cloud provider, with Iceberg-compatible services at every layer of the data stack, so every team can use the right engine for the job without moving data, rebuilding pipelines, or re-implementing governance.
Why Apache Iceberg?
Features like ACID transactions, schema evolution, snapshot isolation, and built-in time travel ensure every query sees a complete, consistent view of your data.
Iceberg allows you to flexibly evolve schemas and change partitioning strategies as your needs change. And because Iceberg is a 100% open-source, community-driven table format, it’s not tied to a single vendor’s roadmap.
Consolidate workloads on a single copy of governed data, eliminate redundant ETL pipelines, and reduce storage costs by removing duplicate datasets across siloed systems.
With native support across popular frameworks like Apache Spark, Apache Flink, and Presto, Iceberg lets you choose the most performant tool for each task.
Benefits of Iceberg on AWS
Apache Iceberg brings reliability and simplicity to your data lakes, but setting up and maintaining Iceberg tables can be an operational tax. Amazon S3 Tables offer purpose-built Iceberg storage that automates routine tasks like compaction, snapshot management, and streamlined governance, returning valuable productivity hours to developer teams.
Learn more about Amazon S3 Tables automated maintenance for Apache Iceberg
AWS offers Iceberg support across the entire data stack, giving you the breadth and depth to choose the right tool for the job. And you never have to sacrifice performance or cost: Iceberg materialized views in AWS Glue accelerate query performance from Apache Spark up to 8x, Amazon EMR runs Apache Iceberg write jobs over 2x faster than open-source equivalents, and Amazon Redshift RG instances run Apache Iceberg workloads up to 2.4x faster than previous generations at up to 30% lower cost.
Learn more about working with Apache Iceberg using AWS engines and services
AWS Glue Data Catalog and Amazon S3 Tables expose Iceberg REST catalog endpoints, so Iceberg tables on S3 work with any compatible catalog. Catalog federation in AWS Glue lets AWS engines like Amazon Redshift, Amazon Athena, and Amazon EMR query tables cataloged in remote Iceberg catalogs without copying data, giving you a unified view of all your data regardless of where it lives.
Learn more about catalog federation to remote Iceberg catalogs
Openness does not come at the expense of governance. AWS allows you to give every engine and user a consistent view of what tables exist, where they live, and how they're structured. With AWS Lake Formation, define fine-grained access policies (table, column, row, and cell level) once in the catalog, and they're enforced consistently across AWS services.
Your data on AWS stays exactly that — yours. Amazon S3 stores data in your own account in open formats accessible by any compatible engine. You know exactly where your data lives, can audit access patterns through detailed request-level logging, and retain complete ownership. Amazon S3 delivers industry-leading durability, and availability, security, and performance while you retain full control and visibility.
Architecture
Amazon MSK Express Brokers now deliver data to Iceberg streaming tables
Continuously deliver real-time data for analytics at lower cost
Learn moreUse cases
Foundation for AI/ML
Give AI models and agents governed, versioned access to production data with lineage, fine-grained access control, and time travel built in. Teams can reproduce training runs, audit agent behavior, and roll back when needed, without building custom infrastructure.
Open lakehouse
Migrate from legacy Hive tables or proprietary formats to AI-ready Iceberg data lakes and lakehouse architectures without downtime or data rewrites. AWS provides in-place migration paths so your teams can modernize incrementally while workloads keep running.
How Yelp modernized its data infrastructure with a streaming lakehouse on AWS
Enterprise scale in-place migration to Apache Iceberg: Implementation guide
Metadata unification and cost reduction
Stop managing metadata across disconnected catalogs. Centralize critical metadata for a trusted, comprehensive view of data assets and eliminate redundant data copies and complex ETL systems.
Access Databricks Unity Catalog data using catalog federation in the AWS Glue Data Catalog
Access Snowflake Horizon Catalog data using catalog federation in the AWS Glue Data Catalog
Petabyte scale analytics
Run interactive queries across petabytes of Iceberg data without moving it into a warehouse. With multiple engines all safely reading from the same tables, analysts and data engineers choose the right engine for each workload without copying data.
Key services
Next steps
Get step-by-step guidance for your Iceberg architecture
Start building with Apache Iceberg on AWS
Did you find what you were looking for today?
Let us know so we can improve the quality of the content on our pages