Page retrieval process of web browsers

If you ask many people who frequently use web browsers only a small portion is able to explain how web browsers retrieve web pages including many other objects like graphics. Therefore, it's time to understand it.

We assume that you sometimes experience that your browser takes some time to render a complete web page. The text is shown first and the graphics arrive later. This already points to a process that seems to consist of more than a single step.

If you enter a URL into your browser and start the request or if you activate a hyperlink, you often ask your browser to request a web page from a remote server somewhere on the Internet. This means, that such a remote server responds with a HTML page. An HTML page consists of a lot of markup code that your browser is able to interpret. You don't need to understand this markup code because this code is important for your browser only and the browser doesn't present it. Besides the markup code, a web page often contains text that your browser presents to you. However, an HTML page doesn't contain graphics but only references to graphics on the same remote server or another server. This is the key to understand the problem. A browser must identify all references to other objects inside an HTML page and request them afterwards. Due to the fact, that most browsers can handle only a limited number of HTTP requests in parallel, a browser cannot request all objects at a time. If an object has arrived, the browser starts the next request for a missing object. This means, that a browser is able to complete a web page after the last object has been arrived and interpreted.

If the communication between a browser and a server takes place via a cellular network with a high latency, the delay between sending a request from a browser and receiving a request at a server and vice versa for the response increases the transmission time for each object.

The more objects a web page contains, the longer it takes to retrieve the whole page including all graphics for instance. So, keep the number of objects as low as possible. Even if a HTML page's volume has a size of a few kilo bytes, it should contain not more than one or two objects. If a web page consists of a Cascading Style Sheet (CSS) and a graphic, the maximum is already reached.

Copyright © 2001-2003 by Rainer Hillebrand and Thomas Wierlemann