For systems that are sensitive to communication latency with the database, the latency caused by inter-AZ communication can be a problem.

In such cases, it will be an option to allocate resources to a single AZ to avoid the latency, at the expense of availability.

However, one thing to note here is that an AZ consists of multiple data centers, so the distance between each EC2 instance and RDS DB instance may vary even within the same AS, which may affect system performance.

This session aims to uncover trends in intra-AZ communication based on actual measurements and provide insights for designing latency-sensitive systems.