Introduction
I am a network consultant. I often help people solve network problems by capturing packets in their network. One day, a customer told me that packet loss exists in their network when using the virtual workspace application Teradici.
They were trying to determine if network problem or hardware shortcomings or server fault lead to packet loss. The customer configured a test bed so that the client was the only one accessing the server, and then captured traffic and sent them to me to analyze.
Teradici doesn’t use TCP (Transmission Control Protocol) where any loss packet could be retransmitted. It uses ESP (Encapsulating Security Payload) which rides directly on top of IP (Internet Protocol). So we can’t use TCP sequence numbers to look for packet loss.
no TCP sequence numbers to use to look for packet loss. But there’s another way, the IP identification number. Each segment passed down from upper layers to IP gets an identification number which shows in the packet. All we have to do is check for skips in the number and we can find loss.
Determine the Level of Packet Loss
Once I had the packet file, I can import packets into Unicorn. About importing packets, please access http://unicorn.net-analyzer.com/en/ . Then switch to packets tab and select a packet item in the packets pane, you will see decoding information of the packet which is showed in left-bottom pane. Next, we find the Identification item in the decoding pane and select it, now you will find that a column of the packets list become Identification, it show Identification value of all packets. Ok, I had Identification value to work. All I had to do is look for any value greater than 1. There were no packets lost on the client side. See figure below
Figure 1: Determine the Level of Packet Loss in packets tab
Summary
It would have been easier with TCP, but more and more PCoIP protocols are in companies’ environments. Now you know it is not impossible to figure out of there is packet loss in the network.