HTTP

Intro

This is a client-server protocol used to send requests and receive responses. It is stateless (the server doesn't remember the client).

HTTP relies on lower levels being reliable

• HTTP uses TCP as its transport protocol, since it ensures packets are always delivered correctly and in order, as opposed to UDP.
• TCP is connection-oriented, i.e. it needs a connection to be established before messages can be sent.

Objects

• An object is a file like a HTML file, PNG file, MP3, etc.
• Each object has a URL.
• The base object of a web page is usually an HTML file that has references to other objects.

URL

A Universal Resource Locator (URL) is used to locate files that exist on servers. They have the following parts:

• Protocol in use
• Hostname of the server
• location of the file
• Arguments to the file

Non-persistent HTTP

Non-persistent connections involve one TCP connection per HTTP request. Once a response has been sent by the server, the HTTP connection is closed.

The main downside is that a whole TCP session has to be built up and torn down every time a single HTTP request is sent. For example, if the client receives a file with several JavaScript elements and images, it will have to open many more TCP connections, which is slow.

Persistent HTTP

This uses a single client-server TCP connection for all the HTTP requests and responses.

HTTP Messages

There are two types of HTTP message: HTTP Requests, and HTTP Responses