Remember FastCGI? (2021)

Summary

“Serverless” is sometimes referred to as “cgi-bin” which isn’t entirely fair as it’s somewhere between cgi-bin and FastCGI. Somewhere along the way both faded from memory. While goofing off last weekend wondered to myself: is FastCGI still useful? Unlike the classic cgi-bin approach where a script or program was executed for each individual request, FastCGI is a binary protocol which allows for longer lived processes serving multiple requests. It continues to be used in the PHP community but seems to have largely fallen out of favor. Nonetheless I decided to tinker a little bit with FastCGI in Rust. The most sensible crate that I found was the fastcgi and played around a bit. The crate is a bit old and I needed to do some fiddling to get a simple example compiling: extern crate fastcgi; use std::io::Write; use std::net::TcpListener; use log::*; fn main() -> std::io::Result<()> { pretty_env_logger::init(); let listener = TcpListener::bind("127.0.0.1:8010")?; info!("Listening on 8010"); fastcgi::run_tcp(|mut req| { info!("Handling request"); for param in req.params() { info!("{:?}", param); } write!(&mut req.stdout(), "Content-Type: text/plain\n\nHello, world!") .unwrap_or(()); }, &listener); Ok(()) } In order to test out my little FastCGI server I spun up nginx in Docker which required a little bit of configuration in the nginx.conf: server { listen 8080; location / { include /etc/nginx/fastcgi_params; fastcgi_pass 127.0.0.1:8010; } } After about 45 minutes of tinkering around I had an end-to-end example working and arrived at the conclusion that if I was going to do all this process management and web server configuration, how is this any better than just running an embedded web server? A functionally identical Rust program using Tide looks something like this: #[async_std::main] async fn main() -> Result<(), std::io::Error> { tide::log::start(); let mut app = tide::new(); app.at("/").get(|_| async { Ok("Hello, world!") }); app.listen("127.0.0.1:8080").await?; Ok(()...