Lecture 15: The wired physical layer

This week, we’ll talk about the lowest layer of the 5-layer stack: the physical layer. This is the layer that deals with the ground realities of transmitting data on a cable or an antenna. It is also the layer that is closer to electrical engineering than computer science. Why are we learning about this layer? It’s not because it is going to be particularly useful in a job that you take up after you graduate. The physical layer is mostly hidden from view and just gets out of the way, a testament to how well it is engineered. Even for many networking researchers, the layer is just a blackbox with a certain functionality. So why learn about it? Opening up the physical layer serves to illustrate an important meta-aspect of computer networks in specific and computer systems in general: you can always open up the blackbox, dig one level deeper, and understand things at one more level of detail. The philosophers call this the infinite regress problem where the validity of any statement can be repeatedly questioned. For this course, however, we’ll stop at the physical layer, and not regress any further. The discussion in this lecture will be quite qualitative. This is not because the physical layer is inherently qualitative. The reality is just the opposite. The physical layer is backed by (in my opinion) some really beautiful mathematical concepts. However, these concepts require a familiarity with proabability, random processes, and signal processing, which I am not assuming for this class. The central question for today’s class is: how do we transmit bits on a wire? Next class, we’ll see how we transmit bits on an antenna and what distinguishes an antenna from a wire. We’ll deal with transmitting bits as electromagnetic (EM) waves—as opposed to (say) sound waves. While there are examples of transmission media that use sound as a communication medium (e.g., SONAR and human communication), the bulk (all?) of digital communication today occurs using electromagnetic waves as the communication medium. So, we’ll focus on this alone. The general concepts discussed here only rely on the “wave” part of EM waves, and they can be generalized to sound waves as well. The major difference is quantitative: communication over a sound medium has lower bandwidth (and hence capacity) than communication over an EM wave medium. We’ll deal with three primary concerns in this lecture: