NGINX is a powerful proxy server for load balancing both of HTTP connections as well as IMAP (and probably many many more).
At cognovís our main usage of NGINX is as a load balancer as well as a domain based proxy server to run multiple installations of OpenACS behind a single IP Address. This should resolve the issue of getting IP Addresses for your server, just because you do not want the user to see the port you are really accessing. 
The typical Nginx Configuration file contains some general information and then for the HTTP access an HTTP section. This http section can be filled with information for the HTTP module and any other of the HTTP* modules listed in the nginx Wiki.
As the typical user you want to define a server description first, 
stating information about your server (e.g. where nginx should listen for this server). Within the server section you will have a new section for each location of the server. This allows you to have a server react to request differently depending on thelocation of the request, e.g. allowing you to serve static files locally from nginx or a different server rather than hitting your main webserver with requests for CSS files as well.