May 28, 2024

TCP/IP and Packets


The original version of TCP was eventually split into two protocols - one (IP) to handle addressing of packets, the other (TCP) to deal with their assembly/disassembly. The design philosophy was based upon the idea that specialised protocols should be responsible for implementing a single task; it was a modular design. A collaboration was required between each protocol.

The layered model of communications

The modular approach provides engineers with the means to represent communications protocols (including Internet protocols) in "layers".
The layers of a TCP/IP stack consist of:

  1. Application Layer - Top of the stack; where the user interacts with the network via the applications. The protocols at this level are embedded in the application programs. Used, for example, to send and receive e-mail (SMTP, POP3), log in to remote machines (Telnet), transfer files (FTP) or browse Web pages (HTTP).
  2. Transport or Host-to-Host Layer - TCP ensures the reliability and integrity of messages and processes them to and from the Application layer above.
  3. Network or Internet Layer - Responsible for ensuring delivery of packets to destination; the role of IP. No guarantee whether packet will be delivered, just decides its address
  4. Link, Physical or Network Access Layer - Lowest level in the stack; responsible for communicating with the hardware which connects a machine to the Internet, e.g. providing a datagram/packet with a Mac address to an Ethernet card linking a machine to an intranet.


Next page » Packet handling

Previous page « Anatomy of a packet





































Up to top of page