Under the hood: Broadcasting live video to millions

This post details the engineering challenges and solutions for scaling live video broadcasting to millions of users. It introduces the 'thundering herd' problem and explains how edge caches, origin caches, and request coalescing are used to manage traffic spikes. It also describes the implementation of RTMP for reducing latency to a few seconds, contrasting it with HLS and highlighting the benefits of its push model and small chunk sizes for interactive live streams.