White paper: Amazon Aurora Improves PostgreSQL Database Performance by 12 X

A leader for a customer-to-customer platform for selling and buying old stuff managed a 6-terabyte database using an early version of Amazon RDS PostgreSQL version 9.3. With the gaining popularity in their platform, they needed a way to scale performance and the size of the database near term to 15-terabytes. The main challenge was that the downtime could not exceed 6 hours. Simple tasks like adding a column to their huge table resulted in locking the table longer than the tolerated downtime.

In this paper, we will walk through options for upgrading their PostgreSQL 9.3 server to 9.6 and then migrating to Amazon Aurora for PostgreSQL. The result was that their functional test script that used to take 2 hours to run with their aging PostgreSQL 9.3 server now runs in 10 minutes on Amazon Aurora!

Now that they are on Amazon Aurora, they now have ready access to all of the proven AWS Machine Learning services without having to build out their own infrastructure or come up with a hybrid solution for moving data to the cloud for processing.