In standard measurements with the IO Graph tool, we measure the performance of the network in units of packets/second, bytes/second, or bits/second. There are some types of data that cannot be measured with these parameters, and this is the reason we have the Advanced... feature in the Y-Axis options.
Choosing the Advanced... feature from the Unit: drop-down menu under Y-Axis opens a wider IO Graphs window, and provides the following options:
To start using the IO Graphs window with the Advanced feature, perform the following steps:
The time delta between frames can influence TCP performance, and there are cases in which we would like to correlate these with the performance we get from the network.
Let's look at the following capture file:
Here, we see packets sent from the source IP 10.2.10.105 as configured in the display filter.
To view the time variance between frames, configure the following parameters:
frame.time_delta
value, configure ip.src == 10.2.10.105
in the field beside Filter: and choose MAX(*) and type frame.time_delta
in the fields beside Calc:frame.time_delta
value, configure ip.src == 10.2.10.105
in the field beside Filter: and choose AVG(*) and type frame.time_delta
in the fields beside Calc:frame.time_delta
value, configure ip.src == 10.2.10.105
in the field beside Filter: and choose MIN(*) and type frame.time_delta
in the fields beside Calc:The graph that we will get is as follows:
What we see in the screenshot is a graph of the minimum, average, and maximum time delta between frames. What do we do with it and how do we use it for network debugging? This will be covered in Chapter 10, HTTP and DNS.
TCP events can be of many types: retransmissions, sliding window events, ACKs (or lack of them), and others. To see the number of TCP events over time, we can use the IO Graph tool with the Advanced... feature and the COUNT(*) parameter.
To do this, perform the following steps:
tcp.analysis.retransmissions
in the filter fieldIn this example, filters were configured to monitor TCP retransmissions on three different TCP streams.
In the graph of the preceding screenshot, you can see that retransmissions from each TCP stream are presented in different colors.
In various network protocols (mostly on those running over TCP), variations in time between frames (that is, the frame-time delta filter) can influence the performance significantly. One of the tools for viewing these changes in the IO Graphs window is the Advanced... configuration.
To do it, perform the following steps:
frame.time_delta
Here is an example. In the following screenshot, we see a packet list with time variations between frames (a second time column was added in order to see the real time and time variations):
You see that there are some large time variations between frames; for example, 29.24 seconds in the frame 1,223, 9.12 seconds in the frame 1,247, and more.
In the IO Graphs window configured as described earlier, you will see the following:
As you see here, there are variations in time between frames. Later in this book, we will learn to see what causes these problems and how to solve them.
The IO Graph tool is one of the strongest and most efficient tools of Wireshark. While the standard IO Graph statistics can be used for basic statistics, the Advanced… feature can be used for in-depth monitoring of response times, TCP analysis of a single stream or several streams, and more.
When we configure a filter on the left, we will filter the traffic between hosts, traffic in a connection, traffic on a server, and so on. The Advanced… feature provides us with more details on traffic. Here are a few examples:
13.58.118.119