Table of Contents Chapter 4 Chapter 6 Netguru.net Glossary
Chapter 5 - Data Communications Technology
We've now had the opportunity to get a rather global view of how data is moved around the various networks out there, both local and wide area. Now let's go beneath the surface and get into the details of data is actually transmitted.
The purpose of this chapter is to clearly explain the differences and characteristics of the two main types of data to be communicated - digital and analog. In addition, the methods for placing these types of data together or separately on a network is discussed.
Analog and Digital Signals
The term "analog" comes from the word "analogous" meaning something is similar to something else. It is used to describe devices that turn the movement or condition of a natural event into similar electronic or mechanical signals. The are numerous examples, but let's look at a couple.
A non-digital watch contains a movement that is constantly active in order to display time, which is also constantly active. Our time is measured in ranges of hours, minutes, seconds, months, years, etc. The display of a watch constantly tracks time within these ranges. In effect the data represented on a watch may have any number of values within a fairly large range. The watch's movement is analogous to the movement of time. In this respect the data produced is analog data.
Another prime example of an analog device is a non-digital thermometer measuring a constantly changing temperature. The action is continuous and the range is not very limited, though sometimes we wish it were. The data produced by a thermometer is analogous to the change in temperature. Therefore, it is an analog signal.
Digital signals, on the other hand, are distinctively different. Digital signals don't have large ranges, nor do they reflect constant activity. Digital signals have very few values. Each signal is unique from a previous digital value and unique from one to come. In effect, a digital signal is a snapshot of a condition and does not represent continual movement.
Of course the most obvious example of digital data is that communicated on-board a computer. Since a computer's memory is simply a series of switches that can either be on or off, digital data directly represents one of these two conditions. We typically represent this on and off status with 1s and 0s where 1 represents an "on" bit and 0 represents "off".
Analog data, by its nature, more closely captures the essence of natural phenomenon, with its action and subtlety. Digital data can only attempt to capture natural phenomenon by "sampling" it at distinct intervals, creating a digital representation composed of 1s and 0s. Obviously, if the interval between samples is too large, the digital representation less accurately represents the phenomenon. If the sampling occurs at too short of an interval, then an inordinate amount of digital resources may be utilized to capture the phenomenon. The changes involved may not be significant enough to warrant so frequent a sampling for accuracy's sake. To digitally represent sound authentically, a sample must be taken over 44, 500 times per second.
A reference to digital resources would certainly include digital storage media. In terms of storage, digital samples of natural phenomenon, or encoding of analog signals from such phenomenon, generally requires a significant amount of recording media (i.e., disk space). To record a second of authentic sound, 1.5 million bits of storage is required. Analog signals don't require such great storage capacity, but they do suffer in the area of duplication.
When copying an analog signal from one generation to another, deterioration of the original signal occurs. A prime example is when we copy a videotape. Since video recorders are analog machines, copying a tape several times results in the accumulation of unwanted analog values called "noise". Eventually these signals become so evident, that the original analog signal is compromised and the video "dub" suffers from intense graininess and poor audio sound. Our technology is limited in the transmission and duplication of analog signals because of the infinite number of values that are allowable.
Digital signals, however, have basically two values. It is much easier to work with two values rather than an infinite number. Consequently our current level of technology allows us to maintain the original quality of a digital signal. With a value of "on" or "off", it's pretty heard to miss.
The Flap Over Digital Music
When digital audio tape machines were trying to get started in the US market, record companies screamed bloody murder. Pirating has always been a problem for the music entities. It's estimated that millions are lost each year on pirated copies of commercially available tunes. Some third world countries have made pirating an art form, releasing exact duplicates of hits from the US with domestically produced cassettes and cassette cases. This profitable business leads to corruption on several layers. I speak from experience having a had a copy of US tunes bound for a friend "disappear" as it passed through customs of a third world country. The only limitation for the pirates (other than laws that are sometimes poorly enforced) is the amount of noise and signal deterioration that accumulates through successive generations eventually destroying quality enough to render the dubs unmarketable.
The obvious danger of digital recording equipment is the lack of deterioration over successive dubs. Consequently, the record companies balked at the notion of having such technology in the hands of the public. Who blames them? The resistance however has not been enough to prevent digital tape technology from being marketed in the US. It is currently available in a variety of formats - which is one of the reasons why it is not used on a widescale basis. There is no clear cut standard that everybody follows without question. In addition, digital to digital copying takes place from one tape to another. However, copying something like a compact disk to tape requires that the digital signal from the compact disk be converted to an analog signal (sound) before going to the Digital Audio Tape (DAT) deck. There the signal is re-encoded into a digital format to be placed on tape. Since the tape holds digital data, there is no way to drop it into a standard cassette player and play it. The standard cassette player plays an analog tape. These restrictions will probably keep the would-be pirates at bay for the time being, but as DAT becomes more widely accepted in the marketplace, the bad guys will undoubtedly rise to meet the demand.
Both analog signals and digital signals have found a home in the networking world. Analog signals are used in a certain type of network known as broadBand networks. Digital data is typically used in what is known as a baseband network.
Broadband networks incorporate technology similar to that of cable television. Data, whether it is video, audio, or digital, is transmitted on the wire at certain frequencies. The typical medium is coaxial cable. Just like you can have cable TV at home bringing you several channels at once, so broadBand systems can bring you several channels of data. More on this type of network will be presented later.
Digital technology is generally utilized exclusively for baseband networks. These networks devote the entire cable (and subsequently its bandwidth) to network transmissions. The baseband network will be more fully discussed later in this chapter.
In comparing analog and digital signals, advantages lie on either end of the spectrum. Analog signals suffer far less from attenuation over long distances. This rather makes sense. Since digital data can only be a 1 or 0, what happens when a signal becomes so weak that it is hard to distinguish between each state? Sometimes we just can't. Analog devices, on the other hand, are equipped to handle the infinite values between 1 and 0.
Digital devices are a lot less sophisticated, meaning that they are fairly easy to manufacture and cost-effective. In addition, digital devices are more resilient to EMI and make more efficient use of the cabling bandwidths than analog systems do.
Converting and Translating Data
Converting analog to digital data, or vice versa, requires special machinery. These devices must be able to capture through sampling the continuous movement of naturally occurring phenomenon as well as reproduce an authentic representation of natural events from digital snapshots. The latter involves the conversion of digital data (1s and 0s) to analog data (like sound).
1. Analog Encoding Methods
Analog data is carried by an alternating current. If we were to graphically represent alternating current, it would appear as a wave, with voltage bouncing above and below the zero level. There are three factors to consider: frequency, amplitude, and phase.
Frequency is the rate at which the current alternates above and below the zero current level. When the current rises above zero, dips below zero and then returns to zero, we say the current has completed one "cycle". The name applied to the number of cycles per second is Hertz (Hz). Therefore, if there are 500 cycles per second for an analog signal, we say the frequency is 500 Hertz (500 Hz).
Amplitude would be viewed as the height (peak) and the depth (trough) of the graphic wave. As analog data travels over distance, the amplitude of the wave decreases. This characteristic is called "attenuation". As mentioned earlier, analog waves are less susceptible to attenuation problems, but occasionally they have to be amplified. The amplitude of analog waves is measured in watts, amps or volts. The measurement decibel is often used to describe the power of a signal. A decibel (dB) allows us to understand the comparison of two different power levels of a signal. For instance, let's say we measured the amplitude of an analog signal about to be sent across a LAN at 600 milliwatts. Now we measure the signal after it has traveled through the network and we find that it has a measurement of only 300 milliwatts. If our measuring device could report in dB, we would find that the change in the signal from source to destination was about -3 dB. The decibel is a measurement of relative change, not actual power. Therefore, the -3 dB change would have been the same even if the original power was 200 watts and the second measurement yielded 100 watts.
Finally, phase describes the difference in the start of the cycle of one signal to the start of the cycle of another. One signal acts as a reference signal, the other signal is the phased signal. A phased signal is created by slightly delaying it in order to cause its peaks and troughs to be out of sync with the reference signal. The level of non-synchronization is measured in degrees. If a signal is 180º out of phase, it means that as the reference signal reaches zero voltage following a peak, the phased signal begins. Thus as the reference signal is peaking, the phased signal is (for lack of a better term) troughing. The figure below illustrates this more clearly.
The importance in looking at frequency, amplitude and phase, lies in the fact that it is these components that can be varied in order to allow an analog signal to carry data. Altering the frequency, amplitude or phase of a signal is called modulation. You see, if we modulate a signal, we make it appear to be different from normal. If we know what the normal signal should be like, then we can compare the normal one with the modulated one. The difference between the two represents the data being carried. We must grasp this concept before we can understand signal conversion. Let's take this one step further by seeing some examples of how data is encoded into analog signals.
AM radio is produced by taking a basic signal (radio wave) and modulating its amplitude according to another signal (i.e. voice and music). AM stands for Amplitude Modulation. We may use the same technology for carrying computer data as well. For digital data, it's called Amplitude-Shift Keying (ASK). Here are a couple of examples:
Figure 5.2: Amplitude Modulation can be used to encode data in analog signals.
FM radio is produced by taking a basic signal (radio wave) and modulating its frequency according to another signal (i.e. music and voice). In this case, FM is an acronym for Frequency Modulation. In the digital data realm, the same technology can be applied using Frequency-Shift Keying (FSK). Figure 5.3 illustrates a couple of examples.
An analog signal may also carry data by having its phase modulated. This technology is used in producing multiple sound channels for motion pictures. Motion picture film has only limited space for carrying sound data. It may not have enough available space to carry the several channels required to create a realistic sound for the audience. By placing several different modulated signals out of phase with one another together, we can in effect carry several different channels of data on one analog signal (carrier). In transmitting digital data, modulating the signal phase is called Phase Shift Keying (PSK). Figure 5.4 provides some phasing examples.
Figure 5.3: Frequency Modulation may be used to encode data into an analog signal as well.
Figure 5.4: Phase Modulation can be used to encode data in an analog signal. The amplitude is varied in some technologies.
As you can see the difference is obvious between the different technologies but the result is the same. Data, whether audio, video or digital, can be encoded and transmitted via analog signals, However, some of these methods offer distinctive advantages over the others.
In analyzing methods for carrying digital data, Amplitude-Shift Keying is fairly easy to accomplish. On the other hand, any kind of amplitude modulated signal is very susceptible to outside interference. This is evident if you have ever tried listening to an AM station during an electrical storm. Each lightning occurrence fires random radio waves through the air resulting in signal interference. The same thing happens to ASK devices. Therefore, ASK is not really suitable for transmission over long distances.
Just as FM radio is not generally affected by weather, neither are FSK transmissions. In spite of this, Frequency-Shift Modulation is seldom used for transmission over high-speed lines as the technology does not allow as many bits per second throughput as PSK does.
Phase-Shift Keying technology is what is utilized by most high speed modern modems. One standard for using PSK is called the Bell 212A specification. It allows four different phases (in degrees) to encode data. The result is a potential for 600 phase shifts per second. Each phase shift represents a certain combination of 2 bits (i.e. 00 01 10 or 11). It then logically follows that since two bits are transmitted per phase shift and there are 600 phase shifts per second, the Bell 212A supports 1200 bits per second throughput.
PSK is very resistant to external interference as it enjoys most of the same characteristics that FM or FSK devices do. The signals encoded using PSK may be used for synchronization purposes as well for the sender and receiver.
2. Digital Encoding Methods
There are numerous methods used to encode digital data directly in digital signals. First, it is important to distinguish digital signals from analog ones. An analog signal, as you recall, is a continuously varying wave. Digital signals simply represent ones or zeros, so they are much less variable than analog. Since digital signals generally only represent one of two values, they are much easier to decode than multi-value analog waves. Plus the lack of multiple values makes digital signals easier to decode even after they have been affected by interference.
Digital signals are used internally in computer devices as well as externally in networks. Earlier, we briefly touched on broadBand and baseband networks. Broadband networks typically use analog signals for transmission, while baseband networks generally use all digital signals.
Digital signals rely on having a reference point on which to build a signal representing a binary digit (1 or 0). If the reference point changes, then distinguishing ones and zeros can be difficult. The reference point is created by grounding. If a network is properly grounded, then data errors are much fewer due to lack of reference voltage problems. Grounding is typically achieved by driving a metal rod several feet into the earth or by attaching the ground wiring to metal piping that is buried in the ground. All excess voltage "drains" off into the ground leaving a voltage considered to be a "zero voltage" or "ground reference voltage".
When Ground Isn't Ground Anymore
Recently there has been much press on grounding problems incurred by networks around the world. The problems, as it has been discovered, often stem from surge protectors used to protect networks. When a surge protector senses a high-voltage spike of electricity, it can shunt the excess voltage to the ground circuit. This allows the potentially dangerous energy to harmlessly dissipate into the Earth. However, it has been shown in testing that the ground circuit itself experiences a temporary rise in voltage. Thus the zero voltage point for the network rises creating less of a distinction between 1s and 0s. The result is data problems. Newer grounding technologies allow handling of surges in different manners other than simply shunting them to ground. The result could save many network administrators a lot of headaches.
A crucial element for digital signals is timing. Timing (or clocking) is used for synchronization so that the communications between two devices can be coordinated. The clocking may be controlled locally on each device after the devices synchronize with one another, or be assisted by "clocking bits", which are special bits used to help synchronize communications. The clocking bits are actually encoded in the information being sent from sender to receiver.
When clocking and synchronization is not used, the communication between sender and receiver is said to be asynchronous communication. Asynchronous communications are slower due to the overhead involved in grouping data together. There must be a logical grouping to separate one byte from another since this is not accomplished via timing.
The various digital encoding methods vary from one another in the manner in which they carry binary data, susceptibility to interference and clocking information. What follows is a few examples of digital encoding methods with descriptions of each:
Uses a positive and negative voltage to represent 1s and 0s. A separate clock signal is used to keep sender and receiver synchronized. Fairly resistant to interference because of the great voltage distance between 1 and 0 signals.
Uses positive voltage but no negative voltage to represent 1s and 0s. Because of low voltage variance, unipolar systems are more given to interference problems. Most use separate clocking signals as polar systems because long streams of 0s and 1s may be confusing if clocks are not exactly synchronized at sender and receiver.
Similar to the other polar methods except that a positive, negative and zero voltage are utilized. Whenever a 1 is encountered, the voltage jumps alternately to the positive voltage or the negative voltage. 0 is always represented by zero voltage. This type of encoding is very resistant to interference.
This coding scheme involves voltage changes midway through the item of digital data being encoded. This serves the dual purpose of providing the type of bit being represented plus providing a synchronization cue for clocking purposes. In this encoding scheme a positive to negative mid-bit voltage transition denotes a 0 and a negative to positive transition denotes a 1. The Manchester coding scheme is known as "biphase".
Similar to the Manchester scheme, this method is also a Manchester biphase encoding scheme. The mid-bit transition from a positive to negative voltage (or vice versa) takes place to offer synchronization (clocking). A 0 is represented by a voltage transition at the beginning of a bit and a 1 is represented by no change in the voltage at the beginning of a bit. This type of coding scheme is implemented in local area networking for token ring systems, while the standard Manchester method is utilized for another popular LAN type called Ethernet.
RZ (Return to Zero) encoding Is a variation on bipolar coding. As in bipolar methods, 1 is represented by a negative voltage while 0 is represented by a positive voltage. However, RZ involves switching mid-bit to zero. This provides clocking information for synchronization and better resistance to interference as it is easy to detect a voltage change mid-bit.
NRZ (or Non-Return to Zero) encoding utilizes transitions between positive and negative voltages to denote 1s and 0s. Transitions are relied on rather than specific voltage levels. A 1 is represented by a transition where 0 is represented by no transition. This method is not self-clocking.
It should be plainly evident now that there is a big difference between digital and analog signal types. The method for encoding signals is also radically different. Each has its own advantages and disadvantages so it is not likely that either is going to completely go away. In real life we often have to convert signals from analog to digital or digital to analog several times before a piece of data reaches its destination. In the next section we'll deal with what's involved to perform conversion from one type of signal to another.
3. Converting Signal Types
Perhaps the most common device associated with signal conversion today is the modem. A modem receives digital data and converts to an analog form for transmission over a media, most typically a phone line. Modem is a shortened form of MOdulator-DEModulator, which means that the device is involved in both creating analog signals from digital data and changing analog data back to digital data (demodulating). Here's how it works:
1. A modem receives its signal from a computer, also known as a DTE (Digital Terminal Equipment).
2. The digital signal is used to modulate an analog carrier signal by either frequency-shift keying or phase-shift keying.
3. The analog signal travels over telephone lines or another medium. Remember analog signals can be broadcast further without attenuation problems.
4. The analog data is detected by another modem which receives and decodes the data on the analog signal.
5. A digital signal is generated by the modem and transmitted to the DTE.
This scaled-down explanation ignores other responsibilities of the modem such as determining at what speed the receiving modem can communicate, detecting carrier signals, dial tones, etc. The official designation for a modem is DCE. This acronym is short for Digital Communication Equipment. The designations of DCE and DTE all fall under a standard known as RS-232-C instituted by the Electrical Industries Association (EIA). This standard governs the way that computers, terminals and modems are hooked up to one another including connections and what pins carry what kind of data.
Please note that Novell materials indicate that the acronym DCE, used above, stands for Digital Circuit-terminating Equipment. This is also correct. If you are preparing for the CNE exams, you may want to commit this other terminology to memory.
Modems are not limited to telephone line use. As mentioned above, other mediums can be used to carry the analog data generated by a modem as well. We have made frequent reference to broadBand networks. Broadband LANs utilize modems to allow several different types of data to share the same piece of cable. Each discrete type of data, whether it is computer data, video, or audio may share the cable because each type of data is transmitted using its own unique frequency. Modems are used to encode data on broadBand LANs in specific frequencies. The method of placing several "channels" of data on a single piece of cable is known as Frequency-Division Multiplexing (FDM) which will be discussed later in this chapter. Modems utilized for broadBand networks are very high-speed modems.
It makes sense to use a modem over long distances, but what about short ones. If we wanted to communicate between two DTEs in the same room, it doesn't make sense for us to hook up modems for them. Since the RS-232-C standard describes which pins carry what data, we can bypass a modem altogether by simply connecting the machines directly to one another. The RS-232 connector has several pins that are numbered. The number 2 pin is typically the one that carries data from the DTE to the DCE (modem). It is known as the transmit pin. The 2 pin of the DTE is connected to the 2 pin of a DCE. The number 2 pin on a modem is known as the receive pin. Knowing this, we can construct a cable that basically connects each DTE's 2 pins to each other's 3 pins to form a connection without a modem. Interestingly enough, such a cable is called a null modem cable.
To send digital data over analog systems we use a modem, but to send analog data over digital systems, we use quite a different device. Today's digitally-oriented businesses are looking for more and more ways to convert traditionally analog data to digital. This makes total sense as digital data is extremely easy to regenerate, is less plagued by interference and is more cost effective. Further, digital data may be directly stored on computer storage media. It's seldom that we pick up a trade mag when some mention of imaging, multimedia or digital voice-mail, isn't made.
This conversion can be performed via a CODEC (COder/DECoder). A codec simply receives an analog signal (such as voice) and samples it. Each sample is converted into a representation of several bits. The bits are transmitted across a digital medium, such as a LAN. As digital data, the voice imprint may be stored, retrieved, filed or digitally compressed or altered. It may be routed across a wide area network link where it will be stored on some other system until someone decides to use it. Then the piece of digital data is routed through another codec where its digital contents are converted back to an analog signal. If it were voice data, then the actual sound of the sampled voice would be heard. If the sampling rate of the original were high enough, the voice would sound quite natural.
Though modems are better known than codecs, the day is coming when perhaps that will change. Most modern sound cards have built-in codecs for performing analog to digital conversion. Sound card sales have increased quite dramatically. Plus, the new standard being implemented for carrying digital data on the public switched telephone network also converts voice data to digital. ISDN is a completely digital system that supports voice lines. Voice recognition technology will also heavily depend on codecs.
Multiple Signal Transmission Schemes
Networks require us to jump through some hoops if we are going to accommodate multiple signals utilizing a single piece of cabling. This need is seen throughout networking whether we are talking about local area networks or wide area ones.
Modern telephone systems must place a large number of calls over a limited amount of bandwidth (i.e. a trunk). Broadband LANs must have several different types of data on a single wire at once. These are examples where "multiplexing" must take place. Multiplexing is the process of putting data from several different sources on the same wire, or, in some cases, putting a large amount of data from a single source on several smaller bandwidth wires. There are several different ways that multiplexing can be accomplished. We'll look at a couple of them.
1. Time-Division Multiplexing (TDM)
TDM is used both in networking and phone systems. It is a process whereby several slower speed signals are divided up and placed on a high speed transmission channel. A multiplexer (also called a "MUX") actually selects which source data will be sent at what amount and places that chunk of data on the wire. It then selects a different source and takes a portion of its data and places it on the wire next. In this manner several "samplings" from several sources can be interleaved on the high-speed communications channel. This can be accomplished because the individual sources are sending their data at a relatively slow speed (i.e. 300 baud), while the outgoing channel has significant speed to accommodate a sampling from each source (i.e. 1200 baud). When the data reaches its destination, another multiplexer disassembles the combination data and places each chunk of data on an appropriate channel to its destination, once again at the slower speed at which it entered the original MUX. Figure 5.5 illustrates the concept of time-division multiplexing.
This same technology is used by phone service providers who must grapple with the task of getting a large number of conversations over limited numbers of wires contained in trunks. If the conversations are broken up and put back together fast enough, no one notices it. For this reason, high speed trunks use time-division multiplexing to carry several conversations at once - and no one is the wiser.
Sampling a conversation of data from several sources may take place on the bit, byte or block level. When only a bit from each source is placed on the wire, we call it "bit interleaving". When a byte is sampled and then placed on a wire with other sampled bytes from other sources, we call it "word interleaving".
MUXs, at both ends of a high-speed link, must synchronize with one another so that the time required for each sampling matches. Otherwise, the demultiplexer would not be able to determine which source signal goes with what destination channel. Timing is obviously an extremely important element to a time-based methodology like TDM.
Figure 5.5: Time-Division Multiplexing allows several devices to share a single medium via interleaving.
One disadvantage of multiplexers that use TDM is that they allocate time for a source's data even if the source is not currently sending any. This is a waste of resources. Special MUXs have been created that only make slots for sources when those sources need to send data. This type of multiplexer must communicate with the MUX at the other end of the link whose data is being sent.
TDM can be used on baseband networks. If you recall, baseband networks only carry one kind of data - digital. Digital data is susceptible to attenuation and interferences. Fortunately, digital data can be used with repeaters that actually regenerate the digital signal and rebroadcasts it at a higher level.
Broadband systems may also use TDM for a particular frequency. The frequencies on a broadBand network are many and varied. They are the product of another type of multiplexing called Frequency-Division Multiplexing (FDM).
2. Frequency-Division Multiplexing (FDM)
FDM allows us to take signals from various sources and place them on a single wire by giving each signal its own frequency. The total bandwidth of the entire cable can be divided up into several smaller bandwidths. These are analog signals that carry data.
The information carried by the analog "carrier" may be encoded using any of the analog encoding methods. Each individual signal source must be routed through a modem. The modem takes the digital data and uses it to modulate an analog signal at a unique frequency. A modem with a different frequency is required for each signal source. A modem must be on the receiving end as well, listening for a unique carrier frequency from the sender.
Figure 5-6: Frequency-Division Multiplexing is used to allow multiple channels of data share a common wire (broadBand networks).
FDM may also be utilized by phone companies who wish to maximize their usage of a limited amount of cable. As mentioned in an earlier chapter, the phone companies typically allow about 4 MHz of bandwidth for calls after filtering.
Broadband networks use technology similar to that of cable TV companies in placing several channels of data on a cable at once. Broadband systems use the different frequencies to separate directional traffic and provide special services. Both analog devices and digital devices can use a broadBand network, but only analog signals are carried on the wire.
Chapter 5 Study Tips
1. Know what an analog signal is and know examples.
2. Know what a digital signal is and give an example.
3. Compare analog and digital signals as far as susceptibility to interference, costs, distance limitations, reliability after amplification.
4. Know the ways a digital signal may be encoded into an analog signal (including ASK, FSK and PSK).
5. Be able to graphically represent ASK, FSK and PSK.
6. Know what comprises a digital signal.
7. Understand the different digital encoding methods.
8. Know what a self-clocking digital encoding method is.
9. Know what MODEM stands for, what the item is, and how it operates.
10. Know about CODECs.
11. Understand TDM and sketch out a diagram to represent it.
12. Understand FDM and sketch out a diagram of its operation.
Table of Contents Chapter 4 Chapter 6 Netguru.net Glossary
Copyright© 1993 & 1997 by Atrium Technical Inc. All rights reserved. Illegal to copy without written permission of ATI.