Changelist

• Added more pictures, Layer 4+ info, higher level layers, better jokes (?)
• Changed O(n) notation to actual numbers

Aliens and turtles invading critical Internet infrastructure. Read more to find out what happens! (Layer 4)

To explain the OSI model, we’ll start from the ground up. How can information from one place to another? Then, we’ll explain why various layers of abstraction are necessary for reliable communication.

The final goal will be reliable and connection-oriented communication across two nodes in a connected graph.

1. Reliable: Messages sent are received exactly in the same order and with the same content.
2. Connection-oriented: Instead of sending messages blindly, send messages and wait for acknowledgement from the receiving side. Upon receiving acknowledgement from the receiving side, the sending side is guaranteed that the receiver has received and processed the message.
3. Two nodes: nodes can be anything. They can be two humans in a social network, or two computers connected to the Internet, or two radios.
4. Connected graph: Node A and Node B are connected in the same graph if $A \leftrightarrow C \leftrightarrow B$ (there exists a Node C connected to both A and B) or if $A \leftrightarrow B$ (A and B are directly connected). In other words, A and B can communicate if A can communicate with C and C can communicate with B.

[transitive closure explanation picture]

These four features will become clearer once we go through some examples. The features also correspond to the guarantees provided by Layer 4 of the OSI model.

Layer 1 - Physical Layer

Information must travel through some physical medium. For example, humans receive information from light (electromagnetic waves medium), sound (air waves medium) or smell (gas molecules medium). When we speak, we transmit information through air pressure waves.

Computers can communicate in the exact same way. They can communicate through sound (https://chrome.google.com/webstore/detail/google-tone/nnckehldicaciogcbchegobnafnjkcne?hl=en) or light (QR codes). Most commonly, they communicate through voltage signals across a copper wire. For example, a microphone can communicate audio waves to a computer by applying a voltage equivalent to the measured air pressure at a specific instant. Do this fast enough, every 1/(44.1 kHz) seconds, and we can get nice sounding audio. Varying the voltage across a continuous range is also called analog communication.

The Internet also works, at the lowest level, by changing voltage across a copper wire. Whereas audio uses analog signals, the Internet works through digital signals. Rather than sending the number “14” by having a voltage of 14, we must send “14” by turning the copper wire “high” or “low” using binary numbers. Limiting voltages to a discrete range has nice advantages (search: digital vs analog communication).

The last example is radio, which doesn’t require wire. Since EM waves, like microwave, IR, or visible, can be generated at once location and received at a nearby location, they seem like a good medium. A wave tells us nothing except for their frequency, magnitude, and existence. Therefore, we can encode information into the amplitude (AM) or frequency (FM). We could also send nothing for a “0” and send something for a “1” (on-off keying). Satellites use more advanced modulation techniques for efficiency.