(The OSI Model, Part 1 may be found here.) Welcome back! Last time we covered the basics of the OSI model and briefly discussed two of the most common forms of serial communications: RS-232 and RS-485. This time we'll take a quick look at TCP/IP and conclude with a look at how something like standard snail mail might look if we press-fit it into an OSI model mold. TCP/IP is just one of a number of protocols that runs on the ubuqitous "Ethernet" interface. Note that the term "Ethernet" refers to a number of protocols and physical layers, but I mention it here because its most common usage describes both a the typical CAT5 cable interface to your PC or router as well as the TCP/IP and UDP protocols. It is this context in which we will use the term Ethernet.
TCP/IP is actually a pairing of two separate layers of the OSI model, but they can be logically grouped into one for the purposes of our example here. Remember, we're keeping it simple for now. In this simplified scenario then, we have a three-step process for putting data on the network. The Application layer is, as before, determined entirely by the specific installation. You may be checking e-mail, perhaps looking up sale prices on Amazon, or checking TalkLikeAMachine for updates. Whatever the data, it is packaged and prepared for transport by the Application layer. The TCP and IP layers, then, handle addressing, packetization and error correction. These steps enable the transmission of large pieces of data (think: streaming video or torrent downloads) as a series of smaller "packets" of data. Each one is assigned a sequence number, a source and destination address (and MAC ID), and a bunch of other information. All this information, when processed on the other end in reverse, allows the destination device to verify both the data itself as well as where it came from The Network Access layer is where "Ethernet" comes in. In the most typical case, this is a CAT5 cable, consisting of 4 twisted pairs of wire (8 conductors total) housed in a flexible jacketed cable, terminated with an RJ-45 jack. On the circuit board this jack is then connected to a pair of small transformers (usually packaged in a single connector referred to as a "mag jack"), and routed to the main processor. Note that in the case of WiFi, only the Network Access layer changes. It uses a wireless radio rather than the Ethernet cable itself, but all other details remain the same for the purposes of this discussion. Just as before, the overall big-picture process is the same. The Application Layer prepares the data, the intermediate layers prepare it for transport, and the final layers actually transmit the electrical signals to send the message. The final example should bring it all home: Physical mail. You've probably used it unless you're young enough to actually believe Pluto isn't a planet. (Sorry kids, there are nine planets and no amount of hashtags will convince me otherwise. Now get off my lawn!)
Traditional "snail mail" involves the same major steps as sending a TCP packet. The contents must be prepared (writing the letter), the mail carrier obtains it and processes the first set of instructions (the written address on the label), the letter is sent to a processing facility where it is addressed again (the bar codes and other markings that appear on the envelope), and finally it's sent to the network (loaded on a truck or plane for long-haul transport). Entire books have been written on the OSI model, its interpretation, implementation and further development. This was intended only as a very brief introduction to the concept, and will serve as background material for future discussions of various protocols and communications methods. You'll hear the terms "Application Layer" and "Physical Layer" thrown around a lot, and it will behoove us to have at least a basic understanding of the concepts as we move forward. I hope this was helpful. As always, feel free to comment below if you wish to share your thoughts. Further Reading: Ethernet, TCP, IP, UDP, OSI Model