Speed Wins: The Basics of how to push more queries through each CPU core

When trying to scale Real-time Analytics projects, we usually focus on things like the infrastructure, the tech stack or the architecture, but we often forget about the basics: the things that can help build highly performant and low-latency queries for each specific use case.

When building SQL queries for billions of rows, you have to unlearn some of the SQL you learnt when you started. You also have to take time to understand the use case and work backwards from it in order to build optimal queries. The effect of this effort is usually underrated but it can boost your project significantly and reduce your infrastructure bill significantly. And all of this simply by learning how to reduce the query time from several seconds or minutes to a few milliseconds, as we will see in a real use case example.

At Tinybird we make it our mission to extract hundreds of queries out of each CPU core per second. And it all starts with the basics. In this talk, we will review the fundamentals of SQL query definition for scalability, and how that leads to a proper data architecture and organization.