Green Unicorn (gunicorn) is an HTTP/WSGI server designed to serve fast clients
 or sleepy applications. That is to say; behind a buffering front-end server
 such as nginx or lighttpd.
 .
  * Optional support for Eventlet, Tornado and Gevent to provide asynchronous
    long-polling ("Comet") connections.
  * Process management: Gunicorn reaps and restarts workers that die.
  * Easy integration with Django and Paster compatible applications (Pylons,
    TurboGears 2, etc.
  * Load balancing via pre-fork and a shared socket
  * Graceful worker process restarts
  * Upgrading without losing connections
  * Decode chunked transfers on-the-fly, allowing upload progress notifications
    or stream-based protocols over HTTP