Case Study: Amazon.com doubles throughput and simplifies scaling with Amazon Web Services (Amazon RDS for PostgreSQL)

Amazon Advertising Leaves Oracle for Simpler Scaling on Amazon RDS

Amazon’s Self-Service Performance Advertising (SSPA) team, which runs Sponsored Products Ads, faced rapidly growing transaction and data volumes that their on‑premises Oracle fleet couldn’t scale to meet. Maintaining and provisioning dozens of Oracle databases created heavy operational overhead—database upgrades and patching consumed 30% of DBA time—and the team anticipated needing more than 30 databases to support future demand.

SSPA migrated the campaign store to PostgreSQL on Amazon RDS using AWS Schema Conversion Tool and AWS Database Migration Service, standardizing the schema and validating performance with a staged, three-step rollout and rollback plan. The move doubled throughput, allowed the team to double the RDS fleet in minutes during peak traffic, cut checkpointing and network load, reduced DBA maintenance burden, avoided a multi‑year rewrite of monitoring tools, and gained automatic storage scaling and cost‑effective pay‑as‑you‑grow capacity.

Amazon.com

Samujjwal Roy

Database Architect Manager