How Phreesia replicated a 30 TB SQL Server database to Amazon S3 with AWS DMS

This article discusses how Phreesia, a healthcare technology company, replicated their 30 TB on-premises SQL Server database to Amazon S3 using AWS Database Migration Service (AWS DMS).

Specifically, the article covers:

• The business challenges that prompted Phreesia to migrate their database to AWS, including data growth, performance issues, and scalability constraints.
• The solution architecture involving AWS DMS, Amazon S3, AWS Glue, and the AWS Glue Data Catalog.
• Considerations for the source SQL Server, such as recovery model, CDC setup, retention period, and LOB column handling.
• Optimizations for the AWS DMS migration, including table partitioning, replication instance sizing, storage provisioning, and monitoring.
• The decision to use the Parquet file format and Iceberg table structure for efficient data storage and handling mutations.
• The benefits and results achieved by Phreesia, such as reduced costs, improved query performance, scalability, and integration with advanced analytics.