The final and topmost layer is the Application Layer. This is where the programmer reigns. There are more protocols used in this layer than I care to mention. Some familiar and some home grown. In simplest terms, data flows from the Application layer of the originating host down through the TCP/IP stack and out the Physical layer across to the Physical layer of the destination host. Once the data enters the destination host's Physical layer, the process is reversed and the data flows up through the TCP/IP stack to the Application layer where it is processed by the application code.

A host could be running multiple UDP or TCP/IP applications at once. With all of that data flying around, one might ask how does the TCP/IP stack know which messages belong to it and where to route the messages when it receives them. Just like UDP, TCP/IP handles these situations by assigning each network connection its own protocol port. A protocol port is actually an internal TCP or UDP address. This address is passed down the stack in the header of each packet of data. IP sends logical host addresses (192.168.0.151) that we humans read in dotted decimal format and TCP and UDP send hexadecimal protocol port addresses (7, 23, 5000,etc.) that we interpret in decimal. You were exposed to IP and port addressing in the UDP chapter. TCP uses port and IP addressing in a similar manner.