HTTP/0.9

Principle

HTTP/0.9 is the first version of HTTP. This version was first written to comply with the exactions given by Tim Berners-Lee about the transport of HTML pages at the CERN. Therefore, he implemented very simple requests, that is to say only the one to get a document (the GET method)! We cannot but ask for a document; it is impossible to send "personal" data to servers.

Why HTTP/0.9? When Tim Berners-Lee invented this protocol, there was no version number. HTTP/0.9 got its number only when HTTP/1.0 was written (HTTP/1.0 is the first HTTP protocol described in a RFC) and it was decided that this new version would be called HTTP/1.0.

Example of HTTP/0.9 request

All HTTP/0.9 requests look like this one:

The requested document arrives straigth after the request has been received, and then the connection is closed by the server.

In HTTP/0.9, there is only the GET method. Everything is performed using this method, even sending data to the server (the requested URI looks then like this: http://www.foo.bar/url?var1=foo; what follows the first question mark means "the variable called var1 is set to 'foo'").

Advantages of HTTP/0.9

HTTP/0.9 has some undeniable advantages: it does not rely on the transport layer (layer 4: TCP or UDP) and it can be used to carry any kind of documents. There is nothing more simple than HTTP/0.9.

Restrictions

HTTP/0.9 has obviously some limitations that will be partially solved by HTTP/1.0, and then HTTP/1.1.

The first drawback is that the connection between the client and the server is closed every time after the server has replied to a request. The consequences are the following ones:

• the client must open a connection for every document to be downloaded, especially for images. With a web page that contains 3 images, the client must open 4 connection in a row, and opening a connection is a slow process.
• the user cannot but wait...
• the network is congestioned by requests to open a connection
• web browser open several connection at the same time (up to 4 for Netscape); servers are then also congestioned.

HTTP/0.9 is also not able to manage caches. Document transfers are not optimised at all.

We can send data to a server only by using a specific GET request, and this limits the amount of data we can send; let's also notice that this data is written in the URI, not hidden or encrypted, so there are problems of confidentiality.

The user is aware of errors (he can see a weird web page), but the web browser does not know there is something wrong happened.

Improvements

It is provided by HTTP/1.0.

Reference

RFC:

• RFC1123: Requirements for Internet Hosts -- Application and Support

 printable format