A plain HTTP server can handle about 25k request per second, while Express and Hapi, can serve respectively up to 8k and 2k request per second.
About a year ago we started a journey with a goal in mind, build the fastest HTTP framework for Nodejs. So Fastify was born, and it can reach and surpass the crazy result of 20k requests per second.
How can Fastify be so fast? In this talk, we will analyze the tools and the techniques we used to optimize it, the cost of abstractions, optimizations and deoptimizations, and the libraries you can use to analyze your code. Furthermore, we will discuss some goodness we give to the user out of the box, such as a plugin system that handles encapsulation and inheritance, enabling a microservices ready ecosystem. What are you waiting? In the time you read this abstract we have already served 600k requests!