Python in a Nutshell, 4th Edition

Chapter 20. Serving HTTP

When a browser (or any other web client) requests a page from a server, the server may return either static or dynamic content. Serving dynamic content involves server-side web programs generating and delivering content on the fly, often based on information stored in a database.

In the early history of the web, the standard for server-side programming was the Common Gateway Interface (CGI), which required the server to run a separate program each time a client requested dynamic content. Process startup time, interpreter initialization, connection to databases, and script initialization add up to measurable overhead; CGI did not scale well.

Nowadays, web servers support many server-specific ways to reduce overhead, serving dynamic content from processes that can serve for several hits rather than starting up a new process per hit. Therefore, we do not cover CGI in this book. To maintain existing CGI programs, or better yet, port them to more modern approaches, consult the online docs (especially PEP 594 for recommendations) and check out the standard library modules cgi (deprecated as of 3.11) and http.cookies.¹

HTTP has become even more fundamental to distributed systems design with the emergence of systems based on microservices, offering a convenient way to transport between processes the JSON content that is frequently used. There are thousands of publicly available HTTP data APIs on the internet. While HTTP’s principles remain almost unchanged since ...