An online retail company with more than 50 million active customers stores purchase data in Amazon S3 and additional customer data in Amazon RDS. The company wants to make all data available for analytics to various teams, with fine-grained access controls and minimal operational overhead. Which solution meets these requirements?

The correct solution leverages AWS Lake Formation to build a data lake, which is ideal for combining diverse datasets (S3 and RDS) for analytics. Lake Formation provides centralized, fine-grained access controls across all registered data sources, including S3 buckets and data accessed via AWS Glue JDBC connections to RDS. This minimizes operational overhead by simplifying security management. Migrating purchase data to RDS would centralize data but RDS is not optimized for large-scale analytics, and its access controls are not designed for data lake-style fine-grained access across different data types. Periodically copying data to S3 with Lambda and querying with Athena, while viable for analytics, relies on S3 policies for access control, which can be less granular and harder to manage consistently across a data lake than Lake Formation. Creating a Redshift cluster involves significant operational overhead for managing the cluster and ETL processes, and Redshift's access controls are specific to the cluster, not a unified data lake.