3

I'm trying to calculate how much bandwidth is utilised for a particular protocol, at the different layers.

Wireshark Protocol Hierarchy gives something like this:

Protocol     %Bytes     Bytes
Frame        100%       158223
 Ethernet    100%       158223
  IPv4       100%       158223
   TCP       100%       158223
    HTTP     32.3%      50786
    SSL      8.03       12708

Now what that mean? Its obvious that all traffic was TCP, and of that, some was HTTP and some SSL.

But what confuses me is that the Byte count for Layers 1, 2, 3 and 4 are all the same. So is 158223 the size of the TCP traffic, or the frame?

I would expect each higher layer to be less, to the value of the header overhead. So something like this:

Protocol     %Bytes     Bytes
 Frame        100%       158223
  Ethernet    100%       157000 (158223 - x per packet, where x is Frame overhead)
   IPv4       100%       154000 (157000 - 18 bytes per packet Ethernet overhead)
    TCP       100%       152000 (154000 - 20 bytes per packet IPv4 overhead)
     HTTP     32.3%      50786
     SSL      8.03       12708

1 Answer 1

4

Generally speaking, everything you capture at L2 is a frame, no matter if it is Ethernet, FDDI, ATM, etc. In your example of the 158223 bytes captured, 100% were Ethernet frames.

100% bytes of those frames contained L3 information. In your example, 100% of them are IPv4 packets. They could also include IPv6 or some other L3 protocol (or contain no L3 information at all).

100% bytes of those packets contain L4 information. In your examples, 100% of them are TCP segments. They could also include UDP for example.

Of those segments, 32.4% bytes are HTTP and bytes 8.03% are SSL.

So, to answer your question, 158223 is the size of both the TCP traffic and the frame traffic (as well as the Ethernet and the IPv4 traffic).

2
  • Ok, but why is each layer not different to each other by the difference of the previous layers header overhead?
    – Yusuf
    Commented Aug 13, 2013 at 6:43
  • Wireshark does not change the captured frames (i.e. removing any headers or encapsulation). The frames are just grouped in this report to allow you to see what kinds of data are captured and what percentage of the traffic falls in each category.
    – YLearn
    Commented Aug 13, 2013 at 7:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.