Untitled Document

An Introduction to the HTTP protocol
by K. Yue, copyright 2000
Revised: September 17, 2000

HTTP connection

A HTTP connection is composed of four steps:
1. The client opens the connection by contacting the server at the URL address
2. The client sends a request to the server with the appropriate URL address.
3. The server sends a response to the client.
4. The connection is closed.
There is only one transaction per connection.
The HTTP protocol is stateless. The server does not store any information of the clients.
To obtain state information from the client, the following methods may be employed:
- Insert hidden fields of the FORM element can be used.
- Use client side cookies.

Client Requests

The client request is composed of the following parts in order:
1. A request header containing request header fields.
  - A method field indicating the method, the identifier of the resource and the version of HTTP used.
  - Accept fields indicating the data types (MIME [Multipurpose Internet Mail Extension] content-type) he client may accept. Note that */* means everything is acceptable to the client.
  - Other fields such as Referrer (come from which URL), If-Modified-Since (whether to use cached copy), User-Agent (information about client).
2. A blank line (CRLF) indicating the end of the request header.
3. The message body (used mainly by the method POST).
The three most commonly used HTTP methods are:
1. GET: To retrieve the URL with user data in the query string of the URL. Used in:
  - ISINDEX: deprecated.
  - ISMAP active imagemap.
  - FORM with the method GET.
2. POST: To retrieve the URL with user data in the message body of the request. Used in FORM with the method POST.
3. HEAD Retrieve the header only from the URL.

Examples:

GET /cgi-bin/formex1?name=Bun+Yue&topic=fish
User-Agent: Mozilla/1.1N (X11; I; IRIX 5.3 IP22)
Accept text/plain
Accept application/html
...
Accept */*

Note that there is a blank line after the last accept field.

POST /cgi-bin/formex2
User-Agent: Mozilla/1.1N (X11; I; IRIX 5.3 IP22)
Accept text/plain
Accept application/html
...
Accept */*

name=Bun+Yue&topic=fish

Server Responses

The server responses are similar in having three parts:
1. A response header with several response fields.
  1. The first line is the status line indicating the version, the status code and the explanation.
  2. Other information about the server and the message such as Date, MIME-version, Server name and version, content-length (of the data portion), content type, last-modified, etc.
2. A blank line indicating the end of the header.
3. The response body which is the html content (such as the content of a html file).

Example:

HTTP/1.0 200 OK
Date: Fri, 04 Oct 1996 18:05:04 GMT
Server: NCSA/1.4.2
MIME-version: 1.0
Content-type: text/html
Last-modified: Tue, 01 Oct, 1996 12:02:03 GMT
Content-Length: 3109

<HTML>
<HEAD>
<TITLE> Dr. Kwok-Bun Yue Home Page </TITLE>
...

HTTP Status Code

The HTTP status code has this general meaning:
- 200-299 Successful transaction.
- 300-399 Redirection: to another location or URL has not changed since last request.
- 400-599 Errors.
The error messages are:

Code	Meaning
400	Wrong request syntax.
401	Authorization required. A list of allowed authorization scheme will also be sent.
402	No Chargeto field on the request for a paid service.
403	Forbidden resource
404	The server cannot find the URL requested.
405	Accessing the resource using a method not allowed.
406	Resource type incompatible with the client.
410	Resource no longer available and no forwarding information exist.
500	The server has encountered an internal error and cannot continue with the request.
501	The server does not support the method of a legal request.
502	Secondary server does not return a valid response.
503	The service is unavailable, because the server is too busy.
504	Secondary server takes too long to respond.