Structural optimization of relational databases

What should you do with your database when you’re facing exponential growth in traffic?

Our project went from 300 RPS to 2000 RPS. In this presentation, we’ll share what has let us keep working downtime-free.

There are standard optimization methods: indexes, caching, and increasing server power. But let’s talk about less-popular ways to modify database structure: denormalization, precalculation, and partitioning. We’ll see where they can help as well as where they might be counterproductive.

While we’ll be using MySQL as an example, the approaches we’ll be talking about are also applicable to other relational databases.