Introduction

One of the most popular ways Linux is used is as a platform for running a web server. These days, most people are familiar with the World Wide Web (WWW). Much like many other distributed systems described in this book, the WWW is built on the client/server model. In the web, clients are the people who "surf," using browsers such as Internet Explorer or Mozilla, generating web requests that are sent to web servers, which respond to these requests. The server is responsible for receiving the request, taking the appropriate actions to find and process the request, and then sending the proper response to the client. Thus, web servers implement the server-side functionality in the WWW and communicate with clients using HyperText Transfer Protocol (HTTP). HTTP is the standard by which clients and servers communicate, allowing interoperability between different vendors and different software. This chapter provides an overview of the following:

• What web servers do

• How web servers use the network

• What steps web servers take to service requests

• What concurrency models are used

• Common tuning options for web servers

• How web server performance is evaluated

We'll focus on how web servers deal with static content, such as HTML files and GIF images. By static content, we mean that the HTTP responses that are provided by the server change relatively infrequentlyfor example, through human intervention. In this context, web servers are similar to file servers in that their main function is to distribute files, albeit files that have special meaning and interpretation to HTTP clients. Web servers also can produce content that is generated more dynamicallynamely, through a parameter-driven program such as CGI or PHP. However, dynamic content generation has evolved considerably beyond simple HTTP. This chapter sticks to relatively simple HTTP requests.