Introduction

A web server is a tool that publishes documents on a network, generally the Internet. HTTP is called a language of the Internet and web servers, apart from browsers, are native speakers of HTTP. Web servers generally listen on one or multiple ports for requests from clients and accept requests in the form of URLs and HTTP headers. On receiving a request, web servers look for the availability of the requested resource and return the contents to the client. The term web server can refer to one or multiple physical servers or a software package, or both of them working together.

Some well known web servers include the Apache web server, Microsoft IIS, and Nginx. Apache web server is the most popular web server package available across platforms such as Windows and Linux. It is an open source project and freely available for commercial use. Nginx, which is again an open source web server project, started to overcome the problems in a high-load environment. Because of its lightweight resource utilization and ability to scale even on minimal hardware, Nginx quickly became a well known name. Nginx offers a free community edition as well as a paid commercial version with added support and extra features. Lastly, Microsoft IIS is a web server specifically designed for Windows servers. Apache still has the major share in the web server market, with Nginx rapidly taking over with some other notable alternatives such as lighttpd and H2O.

Apache is a modularized web server that can be extended by dynamically loading extra modules as and when required. This provides the flexibility to run a bare minimum web server or a fully featured box with modules to support compression, SSL, redirects, language modules, and more. Apache provides multiple connection processing algorithms called multi-processing modules (MPM). It provides an option to create a separate single threaded process for each new request (mpm_prefork), a multi-threaded process that can handle multiple concurrent requests (mpm_worker), or the latest development of mpm_event, which separates the active and idle connections.

Nginx can be considered the next generation of web servers. Its development started to solve the C10k problem, that is, handling ten thousand connections at a time. Apache, being a process-driven model, has some limitations when handling multiple concurrent connections. Nginx took advantage of the event-driven approach with asynchronous, non-blocking connection handling algorithms. A new connection request is handled by a worker process and placed in an event loop where they are continuously checked for events. The events are processed asynchronously. This approach enables Nginx to run with a much lower memory footprint and lower CPU use. It also eliminates the overload of starting a new process for a new connection. A single worker process started by Nginx can handle thousands of concurrent connections.

Tip

It is possible that some terms used throughout this chapter are unknown to you. It is not possible to explain everything in a Cookbook format. A quick Google search for a term will give you more details on them.

Both Apache and Nginx can be configured to process dynamic contents. Apache provides respective language processors such as mod_php and mod_python to process dynamic contents within the worker process itself. Nginx depends on external processors and uses CGI protocols to communicate with external processors. Apache can also be configured to use an external language processor over CGI, but the choice depends on performance and security considerations.

While both Apache and Nginx provide various similar features, they are not entirely interchangeable. Each one has its own pros and cons. Where Nginx excels at serving static contents, Apache performs much better processing dynamic contents. Many web administrators prefer to use Apache and Nginx together.

Note

Nginx is commonly used as a frontend caching/reverse proxy handling client requests and serving static contents, while Apache is used as a backend server processing dynamic contents.

Nginx handles a large number of connections and passes limited requests of dynamic contents to backend Apache servers. This configuration also allows users to scale horizontally by adding multiple backend servers and setting Nginx as a load balancer.

In this chapter, we will be working with both Apache and Nginx servers. We will learn how to set up Apache with PHP as a language for dynamic contents. We will look at some important configurations of Apache. Later, we will set up Nginx with an optional PHP processor, PHP_FPM, and configure Nginx to work as a reverse proxy and load balancer. We will also look at performance and security configurations for both the servers.