Tail latency: When you have a bunch of servers trying to serve a request in parallel, One of the servers might take more time than others which affects the overall response time. In other words, Latency of a tail server(worst performing) is affecting the whole response time. When thinking about scale,distribution and speed. Caching is one of the important thing that stumbles you. Understanding it, mean a lot while you wanted to improve or implement them in other programming languages.