I realize, but that too is a constant per platform. It's about the same amount of effort per web server per ecosystem, since in our we case we reverse proxy to the application server.

If we're talking about shared hosting, that's a whole different ball game.