I am using the formula you see on the attached paper (link: https://drive.google.com/file/d/0B4Ajk8jGD1OxWnBORG9iMG0xSEFudWdpcXd3X2p1Yy1FNkJr/view ) to predict a congestion window (cwnd) from passively collected traffic (link: https://drive.google.com/open?id=0B4Ajk8jGD1OxcVhNdmdaVDQ5Zms). However, I am confused with the value of the MSS (maximum segment size) on the segments where the MSS is NULL (link: https://drive.google.com/file/d/0B4Ajk8jGD1OxSzNYa2k5QUpJbTg/view?usp=sharing). I am using iperf to generate the traffic (sending from a client to a server and capture the traffic on the edge of the network). What is causing for the value of MSS to be NULL?
Link to the full paper: https://drive.google.com/open?id=0B4Ajk8jGD1OxOVNRTGpNUVdyVnZsaGZ6ZzdOc2NUdWZELW13
The formula they have used in the paper is on the second column of page 2. Or is there any other formula to predict cwnd from a passive traffic?
The whole actual trace is found on this link: https://drive.google.com/open?id=0B4Ajk8jGD1OxcVhNdmdaVDQ5Zms
The MSS is a TCP option. The MSS is negotiated early in a connection, so the MSS option only appears in a few packets. The other packets don't include any MSS information.
If you need the MSS to analyze a connection, you will have to capture the initial setup of the connection, and remember the MSS value in the setup. If you don't capture the setup, you will not know the MSS.
answered 15 Feb, 16:56
Guy Harris ♦♦