Quick Start Guide

Wireshark Quickstart Guide
22
Appendix 3: Hits Versus Page Views
It may take more effort than you realize to deliver a web page to
your computer. The first step is to get the raw HTML code for
the page. Getting this code takes several sets of packets – the
details will be left to an exercise to be completed later, but
suffice it to say that retrieval includes setup and control packets
as well as query and response packets. Furthermore, in most
cases the response will be a multi-packet data burst that must be
reassembled into a complete http response.
However, once the page is delivered to the application, the
system has only completed the first step required to display the
web page. Let’s consider a simplified web page in HTML, as
shown in the box below.
<HTML>
<Body>
Look at this pretty Christmas tree.<br>
<img src=tree.jpg>
</Body>
</HTML>
Figure 12: Simplified Web Page
This web page will display a short sentence (Look at this pretty
Christmas tree.), followed by a line break, and then a picture of
a tree. Notice that the picture of the tree is not part of the HTML
page that is delivered. All that gets delivered with the page is a
placeholder that tells the browser to get the picture called
tree.jpg and to put it into a specific spot on the page.
So, once the browser deciphers the web page, it knows it must
make another request of the web server. Now the browser asks
for the picture tree.jpg. As a result, displaying this page takes
two hits on the browser. One hit (or request) was for the original
web page, and the second hit was for the picture to be
embedded into the web page. Each additional picture or external
page element is another hit on the web page.
How many pictures are on a single page? 10? 20? A recent
analysis of the CNN front page indicated over one hundred and
fifty separate files were required to display the page. A lot of
these files are graphic files. This includes tiny graphic arrows,
almost invisible lines, menu choices, and advertisements. In
addition, javascript files, stylesheets, and iFrames can all be
external links, and thus can be additional sources of hits.
This topic is
appropriate for this
guide because it helps
explains the plethora
of packets that add
together to display a
single web page.
However, it is also
interesting to consider
the implications for
the number of ‘hits’ a
web site gets.
Let’s analyze what it
takes to get a million
hits on a web page.
First, assume an
average page has 150
images. In
comparison, this
would be 10% smaller
than CNN’s front
page. Now, assume
each visitor sees three
pages on the web site.
It will take less than
2,300 visitors to get
one million hits on this
hypothetical web site.
150 Hits/Page
X 3 Pages/Visitor
X 2,300 Visitors
= 1,035,000 hits