Frontend Rendering: from server-side pages to hybrid rendering

Over the years the frontend has travelled from the simple request -> HTML -> browser model to systems that use server rendering, static generation, caching, streaming, hydration, edge compute and components executed in different runtimes, all at the same time.

This series follows that evolution chronologically, not as a catalogue of acronyms like CSR, SSR, SSG or ISR, but as a sequence of answers to the problems of the previous architectures.

We start with the classic server-rendered web, then go through AJAX and SPA, the cost of client-side rendering, the return of SSR, static generation and revalidation, all the way to modern models such as streaming, partial rendering, islands, resumability and Server Components.

In every article one question matters most:

what problem were we trying to solve
and what new cost appeared in exchange?

Because most modern rendering techniques do not remove work. They only move it between the build, the server, the cache, the network and the user's CPU.