Theory: Physical Layer#

The first layer in the networking stack is the physical layer, which takes pulses of electricity, light, or radio waves, and translates them into the 1s and 0s computers use to store data. Once the physical layer receives new data, it passes that data up to the next layer in the stack: the data-link layer. In this chapter, we discuss the theory behind how the physical layer does this translation. We’ll look at the different types of media the physical layer might have to deal with and how to transfer one type to another with modulation.

Communication Media#

The channel or system through which a sender transmits a message to a receiver is the communication medium. For example, when we communicate by talking, sound waves are the medium that carries our message. Communications can be one of two forms: digital signals, which can have only one of two values (on or off), or analog signals, which can have a continuous range of values. In general, computers work with digital data, whereas humans and the very physics of the world work with analog data, so we’ll be converting a lot between digital signaling and analog signaling in this layer.

In computer networking, the three types of communication media we most frequently use are:

Electricity transmitted over a wire
Radio waves, either broadcast everywhere with an antenna or directed with a dish
Light transmitted over fiber optics

Let’s consider each of these in more detail.

Wire as a Medium#

When we use wire as a communication medium, we transmit signals by sending electrical pulses along a conductor, which is a material that conducts electricity, like copper. A cable refers to an entire bundle of wires (or fiber optics) and may include several wires, plastic insulation and shielding. It’s not unusual to hear the terms wire and cable used interchangeably, although they’re different.

We need at least two conductors, because electricity works better when it has a closed path for the electricity to flow. This is why batteries have both a positive and negative side: electrons flow out of the negative side and into the positive side.

Fig. 3 shows two ways we can encode the 1s and 0s on a wire.

The electricity turns on for a 1 and turns off for a 0, or the direction of electricity alternates.

Accounting for Problems with Wire as a Medium#

Wire isn’t a perfect conductor; electricity is prone to signal loss and interference, with some energy lost in the form of heat. In addition, electricity flowing through a wire creates a magnetic field, which can transfer the energy to a different conductor or smooth out the signal. This can result in signal loss, also known as attenuation. Magnetic fields from other sources can generate additional electricity in wires, which introduces signals on the wire we don’t want; these signals are known as noise or interference.

To minimize attenuation and interference, engineers have developed a neat trick: first, we start with two wires with electricity flowing in opposite directions (see Fig. 4).

Opposite Flowing Signal — Fig. 4 Two signal wires where the second line has an inverted voltage compared to the first#

When we get interference on our cable, it affects both wires. For example, Fig. 5 shows noise that caused a -1 volt spike in our signal.

Noise In Signal — Fig. 5 Balanced line with noise#

When we receive our signal, we invert the signal on one of the wires (see Fig. 6).

Noise In Signal Inverted — Fig. 6 Balanced line after de-inverting the second line#

Differential Signaling#

We’ve inverted the noise so that it spikes down to 0 volts on one line and up to +2 volts on the other line. When we add the signal on both wires, we get a value of +2 volts—the same value it would have been without the noise. By inverting the noise, we’ve canceled it, as shown in the final signal in Fig. 7. This process is called differential signaling. Unlike this text-book example, real-world noise may not affect both wires exactly the same, so some noise may still be left over. Even if it is not a perfect system, differential signaling significantly reduces the amount of noise we receive.

Signals Combined — Fig. 7 Final signal with noise canceled#

We can further improve our signal by twisting together the two wires in the balanced line to create a twisted pair (TP) and get an even cleaner signal (see Fig. 8).

Twisted Pair Cable — Fig. 8 Twisted pair cable#

Any noise introduced to a twisted pair will be more likely to affect the wires equally. Without twisting, the wire closer to the noise could pick up more noise and preventing us from completely cancelling it out.

Twisting also helps reduce loss of signal by electromagnetic radiation. Electricity traveling through a wire creates a magnetic field, the polarity of which is dependent on the direction of the flow. If we twist a second wire with electricity going in the opposite direction, we set up an opposite field to cancel out the first, which reduces signal loss.

Types of Cables#

The American National Standards Institute (ANSI) and Telecommunications Industry Association (TIA) have put together standards for creating TP cable. You can buy cable from any vendor that meets one of their standards and know that it will meet that standard for limited signal loss. Over time, these standards have evolved. We’ve gone through standards named Level 1, Level 2, Cat 3, and Cat 4 that we no longer use (Cat is short for category). The current standard categories are Cat 5 (100 megabits per second), Cat 5e (1Gb per second), and Cat 6 (10Gb per second). The faster cables have more twists per inch to reliably support the higher speeds.

Note

Over a short distance with good equipment and connectors, you could likely get higher speeds on lower-rated cable. Things aren’t black and white.

The plastic plugs at the end of the cable that makes it easy to hook and unhook your device to the network are called RJ45 connectors. Since the 1930s, analog phones typically used four wires. In the 1970s, customers were allowed to purchase their own phones (before they had to be rented from the phone company), and plugs were standardized using the Registered Jack (RJ) set of specifications—typically, the RJ11 specification for analog voice phones. For the last several decades, networks using TP cable normally have the eight-wire RJ45 plug.

Cat 5 cable has four twisted pairs for a total of eight wires; see The Eight Wires of a Cat 5 Cable for a description of each.

Table 1 The Eight Wires of a Cat 5 Cable#
Wire	Color	PC	Hub
Wire 1	Green Striped	Transmit +	Receive +
Wire 2	Green	Transmit -	Receive -
Wire 3	Orange Striped	Receive +	Transmit +
Wire 4	Blue	Unused	Unused
Wire 5	Blue Striped	Unused	Unused
Wire 6	Orange	Receive -	Transmit -
Wire 7	Brown Striped	Unused	Unused
Wire 8	Brown	Unused	Unused

The Cat 5 standard only uses two of the four pairs of wires: one pair for communication from the computer to the networking hub, another pair for the hub to transmit back to the PC. The remaining wires were meant to be used for old-style analog phone communications. As many office phones now often run on the network using Voice over IP (VoIP), these extra wires are left unused. The more recent standard for Gigabit Ethernet uses Cat 6 cable, and those extra wires are utilized to push through more data by transmitting two bits at a time rather than one.

Ethernet cable can not only carry data, but can also power devices. The Power over Ethernet (PoE) standards use Cat 5 or Cat 6 cable to combine power and data delivery. This can be useful for devices such as security cameras, because only one cable is needed.

Cables can be made out of solid-core wire or stranded wire. Solid-core wire is a single strand of wire of a specified diameter. Stranded wire is made up of many thin strands of wire bundled together to make the desired diameter. Solid-core wire transmits signal better than stranded wire and is usually the better choice. However, stranded wire is more flexible and bends easier, so it can be a great option for short patch cables that you coil and uncoil frequently. You can also buy shielded cable that has foil wrapped around it to reduce interference.

Another popular type of cable is coaxial cable (or coax), which uses a center conductor (wire) surrounded by an insulator, surrounded by an outer conductor (see Fig. 9).

Coax Cable Diagram — Fig. 9 Diagram of coaxial cable. Image by user Tkgd2007, modified by author, distributed under CC-BY 3.0 license.#

Fig. 10 shows a photo of a real coaxial cable.

Coax Cable Photo — Fig. 10 Photo of coaxial cable#

While a twisted pair transmits digital (on/off) signals, coax transmits analog signals with high frequencies; this is the type of signal that we can transmit over radio waves. In fact, coax can carry radio signals directly to an antenna for broadcast. However, some encoding of the original signal onto these high-frequency radio signals needs to be done.

Radio as a Medium#

We can communicate with radio waves based on their frequency, which is determined by how fast the radio wave cycles up and down. We measure frequency in Hertz (Hz), with 1 Hz being one cycle up and down per second (see Fig. 11).

For example, you might tune in to an FM music radio station broadcasting at a frequency of 90.1 MHz (90.1 million cycles per second) or an AM radio at 1040 kHz (1,040 thousand cycles per second); your Wi-Fi might run at 2.5 GHz (2.5 billion cycles per second).

When a radio station broadcasts at 100.1 MHz, it also intentionally broadcasts a bit above and below that frequency. We call this range of frequencies the bandwidth and use this bandwidth to encode our audio data. The radio goes about 90 kHz above and 90 kHz below 100.1 MHz, for a total width of 180 kHz. We’ll give more detail on how this works later in this chapter.

Radio Frequencies and Standards#

Radio waves act differently depending on their frequency. Lower frequencies tend to travel farther than higher frequency waves because they bend and bounce off the atmosphere. It is possible to transmit from one side of the Earth to the other because of this. Lower frequencies also penetrate walls and are harder to block than high frequencies. On the other hand, higher frequencies have more bandwidth for faster data transmission. Because they don’t travel as far, you don’t have to worry about interference from the other side of the globe. The term for how these radio waves travel, bounce, and get absorbed is known as radio wave propagation.

The Institute of Electrical and Electronics Engineers (IEEE) develops a broad array of standards, many of which cover how to create interchangeable electronic parts and devices. The IEEE numbers its standards sequentially, and when it came time to create standards for local area networks, the IEEE used the number 802 as a family to group them. Each standard in the family gets its own working group number; the standard for wired Ethernet is 802.3, and for wireless Ethernet is 802.11. As wireless Ethernet evolves, faster standards are invented. To accommodate these updated standards, IEEE adds letters to the end of the 802.11, such as 802.11b or 802.11n. For computers, the frequencies and standards in the United States are 2.4 GHz (802.11b/g/n/ax and Bluetooth) and 5 GHz (802.11a/h/j/n/ac/ax).

These ranges of frequencies have smaller ranges within them called channels. The 2.4 GHz range goes from 2.401 GHz up to 2.495 GHz. There are 11 channels in the US (and 13 or 14 in many other countries), each separated by about 5 MHz. However, the channel bandwidth is 22 MHz, so the channels overlap.

When setting up wireless routers, you can select the channel you want to use. You can use tools like the Net Analyzer Pro app (see Fig. 12) to find out which routers are running on the same channel.

If you’re having problems with wireless connectivity, it may be because too many other nearby wireless routers are running on the same channel. Try using tools like the Net Analyzer Pro app to choose a channel that has less interference.

Satellite Communications#

Another method of communicating data by radio is satellite communications, where any two points can communicate if they can see any satellite that’s part of a network. With satellite communications, you can communicate from the middle of the ocean, from a rural area, or while travelling where it’s impractical to get a wired connection.

Satellites used for communication often are in geosynchronous orbit, orbiting at exactly the same speed of the Earth’s rotation. To a person on Earth, a satellite in geosynchronous orbit always appears to be in the same spot in the sky, which is useful because you can aim a dish at the satellite and not have to keep moving that dish.

However, there are disadvantages to this. To get a geosynchronous orbit, the satellite must be about 35,786 km (22,236 mi) above sea level. The round-trip time to get there and back is 0.24 seconds, based on the formula 2d/c where d is distance (35,786,000 m) and c is the speed of light (299,792,458 m/s).

This means any time you want to do something over the network, it takes an extra quarter second for the signal to get from you to the server because of the satellite, and it takes an extra half second if you want a reply from that server. This isn’t a problem with types of communication that are mostly one-way, like TV shows, but it is if you have to wait an extra half second for every mouse click or any time you speak into the phone.

One current solution to this delay is having not just one satellite in a geosynchronous orbit, but a whole network of satellites zooming around closer to the Earth in Low Earth Orbit (LEO). Because the satellites are closer, there’s not as much delay. But since the satellites are closer to the Earth, their locations in the sky will change, so you need to launch and maintain a lot more satellites to make sure there is always one above in the sky. This is what SpaceX is trying to do. Starting in 2019 they have been launching thousands of Starlink satellites to provide internet globally, and with only about a 0.03 second delay.

There’s a lot more to radio waves as a medium than I can cover in this book. If you want to learn more, I suggest getting an amateur radio license in your country and finding books that cover projects you can do with radio waves, such as building your own antenna for long-distance communication.

Light as a Medium#

Light can also act as a communication medium for computer networking. We can use several different kinds of light to communicate, such as Infra-Red (IR) for remote control devices.

Fiber optics provide a few advantages over wire: they don’t lose signal by electromagnetic loss, nor do they pick up noise from stray magnetic fields. Because of this, they can send a signal farther. Bundling multiple fibers together into one cable requires very little extra space compared to copper wire. Fiber optics do require more electronic components to send or receive the signal, which makes them more costly than copper wire, so if you don’t need the additional speed and distance, you can save money and go with wire.

How far fiber sends a signal depends on the speed, quality of cable, and quality of equipment. It’s common to get a fast data connection with fiber that goes a mile (1.6 km), while Cat 6 cable limits us to around 300 feet. With higher quality fiber and electronics, it’s possible to get a decent signal at runs of about 50 miles (about 80 km), which is particularly useful for undersea cables.

Note

To learn more about the undersea fiber optic cables in use, check out this fascinating website: Submarine Cable Map. If you want to know more about how the backbone of the US internet is put together on land, consider reading this paper: InterTubes:AStudyoftheUSLong-haulFiber-optic Infrastructure

Types of Communicating#

A channel of communication can send a signal in only one direction at any given time. To send and receive signals at the same time, you need two channels, whether it be two sets of wires, two different fibers, or two different radio channels. Given this, there are generally three types of communication:

Simplex#

transmits in one direction using one channel, such as broadcast radio.

Half duplex#

transmits in two directions but uses only one channel, preventing you from talking and listening at the same time. Think old-style walkie-talkies or CB radio.

Full duplex#

transmits in two directions using two channels, letting you talk and listen at the same time, like talking on the phone.

Types of Modulation#

When communicating, we often need to take data or a signal and put it on a different medium in a process called modulation. For radio, we take an audio signal and modulate it to a radio signal. The process of taking the data or signal off that new medium and back into the original signal is called demodulation. A device that does both of those is called a modulator/demodulator, or modem for short. For many years, digital communication was done with a modem over phone lines. Now modems are used for cable, DSL, and fiber optics as well. Because these modern modems are often part of devices that also serve as network routers and Wi-Fi access points, they go by various names.

Radio Modulation#

For music and voice radio, we need to take an audio signal and modulate it to a radio signal. We can hear audio signals from about 20 Hz to 20 kHz, but we need to modulate this audio signal to a radio frequency of 1,000 kHz or 100 MHz. The following are some forms of radio modulation that let us accomplish this.

Amplitude Modulation#

Amplitude Modulation (AM) takes a lower frequency audio signal, shown at the top of Fig. 13, and combines it with a higher frequency radio signal (such as 1,040 kHz), which we call the carrier frequency, shown in the middle of Fig. 13.

To carry our audio, we vary (modulate) the amplitude of the carrier signal so that we can combine it with the audio signal. When the audio signal is at the low part of the wave, we don’t transmit the carrier signal, resulting in a wave with an amplitude of 0. When the audio signal is at the height of the wave, we transmit the carrier signal at full strength, resulting in a wave with a high amplitude. This leaves us with our resulting Amplitude Modulated signal, shown at the bottom of Figure 2-11.

AM was the earliest form of modulation invented for transmitting voice by radio and is still used today in AM radio. Unfortunately, AM wastes a lot of power, and rather than ignoring any electromagnetic noise from sources like old electric motors or storm lightning, they are decoded as pops and crackles.

Frequency Modulation#

Frequency Modulation (FM) changes the frequency of the carrier, rather than the amplitude, based on the signal we want to modulate (see Fig. 14).

Our carrier frequency might go between 90.0 MHz to 90.2 MHz. With FM, if we get spikes from extra radio noise, it doesn’t come across as noise on our speakers, since the noise affects the amplitude, not the frequency. Although FM is noise-free, it takes more bandwidth. AM radio channels take up 10 kHz each, while FM radio channels take 200 kHz. FM is only appropriate to use at higher frequencies where more bandwidth exists.

Analog and Digital Modulation#

In addition to modulating radio signals to work for us, we can also modulate computer signals. Radio works by controlling high frequency electromagnetic waves. Computers work by using 1s and 0s. To move between analog signals like our voice and the binary on/off signals that computers use, we use two methods: pulse code modulation and pulse width modulation.

Pulse Code Modulation#

Pulse code modulation (PCM) allows us to take an analog signal with a range of values and encode it for a computer that runs digitally with only 1s and 0s, and vice versa. Any time we store an analog signal, such as voice or music on a computer, we use PCM to do an analog to digital (AD) conversion. When we play back audio, we use PCM to do a digital to analog (DA) conversion.

PCM samples an analog frequency at regular intervals and converts the current analog signal of each sample into a number. The sample frequency is how often we sample the analog frequency, and the sample depth is the potential range of the converted number. A sample depth of 8 bits (eight 1s or 0s) has 2⁸ (256) different levels going from –128 to +127. Increasing the sample depth to 12 bits would give 2¹² (4,096) levels going from –2,048 to +2,047.

A CD uses a sample frequency of 44.1 kHz and a sample depth of 16 bits, so in one second, the PCM stream records 44,100 numbers that can span 2¹⁶ (65,536) different levels from –32,768 to +32,767.

Fig. 15 shows an example of taking an analog wave and sampling it with PCM.

PCM Sampling — Fig. 15 PCM sample (stepped wave) of an analog signal (smooth wave)#

The resulting wave looks like Fig. 16.

PCM Result Wave — Fig. 16 Resulting PCM wave only#

The faster the sample frequency, the higher the frequency we can store. The higher the sample depth, the more accurately we can represent the signal. Fig. 17 shows how much closer to the original the wave looks if you double both the frequency and sample depth, making it more accurate to the original signal.

PCM Result Wave Double-Sampled — Fig. 17 Resulting PCM wave with double the sample frequency and double the sample depth#

The better the PCM signal, the more data it takes to store it, so Fig. 17 takes four times the data storage that Fig. 16 does. In general, PCM takes so much data that streaming and downloading music wasn’t reasonable until data compression algorithms like MP3 came along.

Pulse Width Modulation#

Another way to represent an analog signal with a digital one is to use pulse width modulation (PWM).

PWM is an efficient way to manipulate the brightness of lights or the speed of a motor. To dim an LED light, rather than turning it on 100 percent of the time or changing the voltage going to the light, we very quickly turn it on and off at a certain proportion. We can dim the light by half by turning it on and off incredibly fast, so that it’s on 50 percent of the time and off 50 percent of the time. For this proportion of on and off, we say that the LED has a duty cycle of 50 percent. Fig. 18 shows a diagram of PWM.

PWM signals allow us to drive motors and use lights efficiently. A light running at 75 percent uses only 75 percent of the electricity. We also use PWM signals to drive servos, which are robotic motors whose position we can control. Robotics, drones, and other automation projects make heavy use of PWM.

Clock and Data Lines#

One of the simplest communication methods at the physical layer is to transfer 1s and 0s from one spot to another by turning the voltage on a wire on or off. If we stream the bits one after another, we call it a serial communication line. Streaming multiple bits in parallel, each on their own wire, is called a parallel communication line. Serial communication is more common than parallel communication, as fewer wires in a cable makes it cheaper.

We can move the 1s and 0s across a wire (called the data line) by rapidly switching the voltage on or off. A positive voltage is a 1; no voltage is a 0. The two most common standard voltage levels for a 1 are +5v and +3.3v. You can convert between the two standards using a level shifter.

In addition to the data line, several serial communication protocols use a serial clock line. The serial clock (SCLK) line tells us when to read the data line by keeping a common clock between both devices. The clock regularly alternates between on and off. Each time the clock voltage falls from positive to no voltage, that tells the receiving computer to read from the data line. If the data line is positive, we have a 1. If the data line has no voltage, we have a 0.

Fig. 19 shows a serial clock line working in conjunction with a communication line transmitting data.

Fig. 19 Encoding data with a serial clock line#

On the rising edge of the clock signal, the data line voltage transitions to either high or low, depending on the data. We don’t read the data during the rising edge of the clock because the data is still transitioning. When the clock signal falls, we hold the data signal at the proper value. The receiver reads from the data line whenever the clock signal is in a falling state.

The clock line is necessary. If our message is a long string of 0s, the receiver can’t tell the difference between that and a wire that isn’t hooked up to anything. If we send 1,000 1s in a row and the clock on one side is slightly faster, we might get 999 1s or 1,001 1s. With high-speed communications, the clocks have to be synchronized exactly. By sharing a clock on its own wire, we solve the issue of synchronizing when to read the data.

Each 1 or 0 is a bit. Computers work with bits in groups of eight. A group of eight bits is a byte, which can store 2⁸ (256) different combinations. We’ll describe how to work with individual bits coming from the physical layer and grouping them into bytes in the tutorial on bit shifting in Chapter 3.

The physical layer is all about how to take pulses of electricity, light, or radio and convert them back and forth between 1s and 0s. This method of encoding the 1s and 0s onto a couple wires is simple and works well; we’ll use it in the project for Chapter 3.

Manchester Encoding#

It’s possible for one wire to fulfill the functions of both a clock line and a data line. Engineers figured out how to do this with Manchester code, a self-clocking signal that has regular transitions between on and off regardless of what we’re transmitting. A long string of 1s or 0s won’t throw off the receiver. We can use those transitions to synchronize when to read rather than using a separate wire with a clock signal, as shown in Figure 2-18.

Fig. 20 Manchester encoding#

Manchester encoding divides each transmitted bit of data into a frame of time. In the middle of that frame, we transition from low to high if the bit is a 1, or from high to low if the bit is a 0. The transition signifies whether we have a 1 or 0, rather than whether the line has a voltage on it. Because every bit that transmits has a regular transition, we don’t need a separate clock line, since the transition itself also acts as the clock.

At the beginning of the frame, we may or may not transition between low and high depending on whether we need to transition to low or high to get ready for the upcoming mid-transition. The pseudocode for the logic used is as follows:

Listing 1 Manchester Encoding#

 if clock signal is rising:
     if data = 1:
         transition low to high
     else if data = 0:
         transition high to low
 else if clock signal is falling:
     if data = 1 and current signal is high:
         transition high to low
     else if data = 0 and current signal is low:
         transition low to high
     else:
         do nothing

Because of the regular transitions, we don’t have to worry about getting desynced by a long series of 1s and 0s. In the next chapter, you’ll implement this code and use Manchester encoding to pass data between two computers.

Review#

A communication medium is the means by which we move a signal, usually electricity on a wire, radio waves, or light on a fiber. The physical layer is concerned with transmitting 1s and 0s from our source via the medium and pulling them off once the signal arrives at its destination.

Signals can attenuate (or lose strength) as they travel over the medium and pick up unwanted interference along the way. Using differential signaling and twisted pair wire helps send signals more reliably and over longer distances using wire. Fiber optics transmit signals by light; they can transmit data farther and faster than wire, but at greater cost.

Radio waves can transmit a signal without the use of cables. The greater range of frequencies used, the more data can be transmitted. Wi-Fi has evolved through several different versions under the IEEE 802.11 set of standards. Each of the evolving standards are a step forward technology, supporting faster and more reliable connections.

There are many ways of encoding signals and converting audio or digital data so that we can transmit them through our desired medium. AM and FM are commonly used for transmitting voice and music over radio. PCM converts audio and music to the 1s and 0s that computers use. PWM controls motors, servos, and lights. We use standards like using clock and data lines and Manchester encoding to transmit digital data from one spot to another.

In the next chapter, you’ll get hands-on experience with signaling through creating your own implementation of the physical layer by directly controlling pulses of electricity with a Raspberry Pi computer.