Suppose we initiate Upload Ping from Client PC to Application Server.
Below is how ping request goes to Application Server: 1. Prior to Ping, script adds a route on Client PC, which ensures Ping request to 192.168.0.125/70 will go through air interface. 2. Client PC sends a ping request, identified with destination IP 192.168.0.125/70. 3. UE forwards ping request to landslide (192.168.0.108) via air interface. 4. Landslide forwards it to switch (192.168.0.1), and switch finds192.168.0.125/70 is within its subnet, thus forwards the request to Application Server.
Below is how ping response goes to laptop: 1. Application Server sends ping response, identified with UE IP 192.168.7.x (CDMA) or 192.168.9.x (LTE). 2. Ping response goes to switch (192.168.0.1). Switch is responsible for data packets within same subnet(192.168.0.x subnet), here switch figures out IP 192.168.7/9.x is not in the same subnet, thus switch will forward the data packets to the router. 3. Router checks its internal IP route table, found “ip route 192.168.7/9.0 255.255.255.0 192.168.0.108”. This means ping response to 192.168.7/9.x will go through 192.168.0.108 (landslide IP). Thus router forwards the ping response to landslide. 4. Landslide forwards the ping response to 192.168.7/9.x (UE) through air interface, thus laptop gets ping response.
We can capture Wireshark logs on each node, and then investigate which node lost packets following the logic above.