You could summarize key aspects of the online operation of most organizations using a few concepts: data access, management, analysis, and delivery. For many of the businesses that turn one or more of those big-data concepts into revenue, PostgreSQL provides an enterprise-level way to handle the complex database workloads that make it all possible, including online transaction processing (OLTP). End users expect fast, readily available OLTP interactions backed by highly available systems. Cloud-hosted database environments can support the performance goals, but enabling high availability can force companies into an uncomfortable tradeoff between redundancy and application performance. For organizations weighing which cloud platform to trust with their production PostgreSQL work, real-world performance data, measured both with and without high availability, can help them make the right call. 

We benchmarked two comparable cloud database environments running an OLTP workload: an Azure Database for PostgreSQL – Flexible Server instance with Dv6-series compute and Premium SSD v2 storage, and an Amazon Relational Database Service (RDS) for PostgreSQL instance on an m7i configuration with GP3 storage. Both instances had 16 vCPUs. We used the industry-standard HammerDB TPROC-C workload, which simulates a demanding warehouse transaction system, to measure throughput in new orders per minute (NOPM). The Azure instance was the clear winner in each case we tested, handling 72.7 percent more NOPM than the Amazon RDS instance without high availability and 79.2 percent more with high availability.  

Our findings show that Azure Database for PostgreSQL – Flexible Server can deliver a large OLTP throughput advantage over Amazon RDS, with or without high availability. That combination of reliability and performance means that organizations can satisfy availability-dependent SLAs without sacrificing end-user experience.  

To dig into the details of our cloud-hosted PostgreSQL comparison tests and conclusions, check out the report below.