The Open Systems Interconnection (or OSI), model of networking is, perhaps, not the most exciting part of the Internet. Some people even consider it dry and theoretical. However, it’s six shades of useful, so we’re going to talk about it anyway. The OSI model descirbes seven layers in networking, and some of us remember using the mnemonic “All People Seem To Need Data Processing.” The actual names of those layers are Application, Presentation, Session, Transport, Data link, and Physical. This overview, which doesn’t include the space to explain all terms, should still be helpful for a beginner to begin to get a better understanding of OSI.
The physical layer (layer 1) is everything you can touch with your hands, especially cabling and connectors. Layer 1 is about getting the signal from one end of the network to another. Layer 2, the Data Link layer, is where the basic protocols of device communication happen, and it’s the place where software is given a chance to detect hardware problems. Data link includes protocols such as PPP, Point to Point Protocol.
Layer 3, the Network layer, is where routing happens, and packets are sent in the direction of the MAC address they should be headed off to. Layer 3 includes all routing protocols. Layer 4, the Transport layer, parses data and puts it into frames to transport across the network; this is where TCP and UDP come into play. Layer 5 is the Session layer. This is where communication between devices is managed; for instance, when you ask Google for information, first a session must be opened between you and the Google server. A session may be half-duplex, full-duplex, or simplex. Layer 6, the Presentation layer, is where you find things like character encoding and application encryption and decryption such as SSL. Layer 6 and Layer 7 are similar, but distinct; Layer 7 (the Application layer) is the actual application that you see and use on the screen.