LLNG is able to use different web servers to provide its services. Here is a resume of all possibilities. We recommend:
Since 2.0, both portal and manager are native FastCGI applications. They can be used on any web server that can dial with a FastCGI server. Some examples:
Application protection overview
Applications can be protected:
by a LLNG handler
by themselves if they can dial with a supported protocol (SAML
To protect applications with handler, LLNG can be used in two mode:
Direct Application Mode : LLNG handler is an embedded application. Handler must be installed on application Web Server
ReverseProxy Mode : applications are hidden behind a ReverseProxy which provides the required LLNG handler
Direct Application Mode
LLNG handlers can be installed on the following web servers:
External servers for Nginx
Natively, Nginx supportes FastCGI and uWSGI protocoles.
Therefore, LLNG services can be provided by compatible external servers.
FastCGI or uWSGI server(s) can be installed on separate hosts. Also you can imagine a global cloud-FastCGI/uWSGI-service for all your Nginx servers. See more at SSO as a service (SSOaaS)
By default, LLNG provides a Plack based FastCGI server able to afford all LLNG services using FCGI engine.
However, you can use some other FastCGI server engines:
(*) LLNG Node.js handler can only be used as Nginx `auth_request` server, not to serve Portal or Manager