|
|
|
|
Concepts Reinforced
OBJECTIVES Upon successful completion of this chapter, you should be able to:
IntroductionAs new vocabulary and concepts are introduced in this chapter, it's important to understand that these words and concepts will be used throughout your study of data communications. Future concepts will be built upon the foundation of a thorough understanding of the vocabulary in this chapter.This is the time to begin to speak to this foreign language which we call data communications. Don't be timid. Try to get together with other data communications students or professionals outside of class and practice your new vocabulary. Discussions of specific technology will be kept to a minimum in this chapter, concentrating instead on the conceptual aspects of the "how" of data transmission. Although modems will be mentioned in this chapter, the in-depth study of technology, including its business impact, will begin in Chapter 3. End-to-end data communicationsFigure 2-1 serves as a roadmap for all of the concepts to be studied in Chapter 2. An end-to-end communication session between two PCs (personal computers) and their associated modems via a dial-up phone line offers an overall scenario into which most data communication concepts can be introduced. In this manner, each concept can be understood individually as well as in terms of its contribution to the overall end-to-end data communication session.Another reason for choosing a modem-based communication session as a means of introducing basic data communications concepts is the likelihood of encountering modems and dial-up connections in real life. This contributes to familiarity with the process as we begin this study, and a greater understanding of this common data communications opportunity at the conclusion of the chapter. However, it is important to understand that the concepts outlined on Figure 2-1 and explained throughout the chapter are not specific to only PCs or modems. These are basic data communications concepts that form the basis of the vast majority of data communications technology. Likewise, although PC modems are discussed in this chapter, modems also have many other applications such as communicating with mainframe computers as described in Chapter 9. Figure 2-1 Basic data communications concepts: end-to-end modem-based data communications session The end-to-end data communications session between two PCs and its constituent concepts and processes as illustrated in Figure 2-1 can be logically sub-divided into three major sections:
In turn, each of these major sub-sections is divided into several constituent processes. Each of these processes will be described conceptually. In addition, any significant protocols or standards currently governing these processes, as listed in Figure 2-1, will be explored. More detailed information on the operation of the carrier network and services will be provided in:
Computer to ModemCharacter EncodingIn order to get data from its source PC to its destination PC, it must be transformed from its humanly understandable form (letters, numbers, voices, images) to an electronically-based machine understandable form. The process of transforming humanly readable characters into machine readable code is known as character encoding.Using a particular encoding scheme, characters are turned into a series of ones and zeroes. Why ones and zeroes? The one and zero are used as symbols to represent two discrete states, much like a light switch being on or off. These discrete states can be easily represented electrically by discrete voltages of electricity. In turn, these discrete voltages of electricity representing coded characters can then be easily transmitted, received, and examined by data communications equipment. The individual 1's and 0's which constitute a given character are known as bits. The series (usually 8 ) of bits representing the entire encoded letter is known as a byte. These 1's and 0's , or digits, represented by discrete voltages of electricity are in digital format and known as digital data. Now that the humanly-readable character is in machine readable form, it (these bits) can now be transmitted in a process generally known as data communications. Characters can be encoded according to a variety of protocols or standards. A few of the currently popular encoding standards are described in the following paragraphs. ASCIIAmerican Standard Code for Information Interchange (ASCII) is one standardized method for encoding humanly readable characters. ASCII uses a series of 7 bits to represent 128 (27= 128) different characters including upper and lower case letters, numerals, punctuation and symbols, and specialized control characters. One use of control characters will be explained further in Chapter 3. An eight bit, known as a parity bit, is added to 7-bit ASCII for error detection. Error detection and parity checking will also be described in Chapter 3. Figure 2-2 is an ASCII table.
Figure 2-2 ASCII Table
Using ASCII and EBCDIC TablesUsing ASCII or EBCDIC tables to interpret character encoding is relatively straightforward. The tables are arranged according to groups of bits otherwise known as bit patterns. The bit patterns are divided into groups. In the case of ASCII, bits 6 through 4 are known as the most significant bits (MSB) while bits 3 through 0 are known as the least significant bits (LSB). In the case of EBCDIC, bits 0 through 3 are known as the MSB and bits 4 through 7 are known as the LSB.In order to find the bit pattern of a particular character, one needs to just combine the bit patterns which intersect in the table at the character in question, remembering that most significant bits always come before least significant bits. In the case of ASCII, this means that bits are arranged from bit 6 to bit 0, while EBCDIC is arranged from bit 0 to bit 7. As an example, representative characters, numerals, and control characters and their bit patterns are highlighted with shading in the ASCII and EBCDIC tables and are displayed in Figure 2-4 in humanly readable, ASCII, and EBCDIC formats.
Figure 2-4 Humanly readable, ASCII, and EBCDIC coding UNICODE and ISO 10646ASCII and EBCDIC coding schemes have sufficient capacity to represent letters and characters familiar to people whose alphabets use the letters A, B, C etc. However, what happens if the computer needs to support communication in Chinese or Arabic? It should be obvious that 128 or 256 possible characters will not suffice when other languages and alphabets are considered.In view of this fact, an international effort was undertaken to establish a new coding standard which could support many more alphabets and symbols than ASCII or EBCDIC. Although Unicode and ISO (International Standards Organization) 10646 were initiated separately, Unicode Version 1.1 and ISO 10646 are identical and were released in 1993. Unicode is a 16 bit code supporting up to 65,536 possible characters (216= 65,536). It is backward compatible with ASCII as the first 128 Unicode characters are identical to the ASCII table. In addition, Unicode includes over 2000 Han characters for languages such as Chinese, Japanese, and Korean. It also includes Hebrew, Greek, Russian, and Sanskrit alphabets as well as mathematical and technical symbols, publishing symbols, geometric shapes, and punctuation marks. Application programs which display text on a monitor must encode characters according to an encoding scheme understood by the computer's operating system. It is up to the operating systems vendors to include support for particular encoding schemes such as Unicode/ISO 10646. Microsoft's Windows NT is one example of an operating system which supports Unicode. Serial vs. Parallel TransmissionThese bits which represent humanly-readable characters can be transmitted in either of two basic transmission methodologies. They can be transmitted either simultaneously (parallel transmission) or in a linear fashion, one after the other (serial transmission). The advantages, limitations and typical applications of each transmission methodology are summarized in Figure 2-5 and illustrated in Figure 2-6.
Figure 2-6 Serial and parallel transmission illustrated
Physical Interfaces vs. Transmission Protocols
With either parallel or serial transmission, it is important to
distinguish between those standards which describe the connectors
or physical interfaces which are used to connect appropriate cables
to a computer's parallel or serial ports, and the standards which
describe the electrical characteristics, or transmission protocol
of either serial or parallel transmission. These differences in
standards are highlighted in the sections on serial and parallel
transmission which follow.
Figure 2-7 Serial transmission physical interfaces
It is important to note that the designators DB-25, DB-9, and
M Block only describe the physical connectors and do not imply
anything regarding the transmission protocol which defines the
electrical specifications for transmission using one of these
physical interfaces. As will be seen, these physical interfaces
can be employed using a variety of different transmission protocols.
Although all 25 pins are defined, in most cases, ten or fewer
of the pins are actually used in the majority of serial transmission
applications. On some PC's, such as personal computers as well
as many notebook and laptop computers, the serial port has only
9 pins (DB-9 connector) and the RS-232 serial transmission protocol
is supported as listed in Figure 2-8.
Figure 2-8 RS-232 Serial Transmission Protocol as defined
for DB-25 and DB-9 connectors
Recalling that character encoding ensures that all characters
can be represented as a series of 1s and 0s, it is the job of
the transmission protocol to represent these 1s and 0s as discrete
electrical signals. RS-232 defines voltages of between +5 and
+15 volts DC on a given pin to represent a logical zero, otherwise
known as a space, and voltages of between -5 and -15 volts DC
to represent logical ones, otherwise known as a mark.
The cable has several small insulated wires within an outer jacket.
These cables come with different numbers of "inner"
wires depending on how many signals need to be transferred from
one serial port to another. Each signal to be carried, or RS-232
pin to be supported, requires its own individual inner wire.
The next question is: Which of the possible 25 signals are most
meaningful, and therefore worth transferring over the modem cable
in this example? Figure 2-9 summarizes the 12 signals which are
most commonly included in modem cables and designates which signals
are assigned to which pins on both DB-9 and DB-25 connectors.
The RS-232 signals described in Figure 2-9 are arranged in logical
pairs in order to increase understanding rather than in order
of DB-25 pin number.
The modem cables will have 12 data leads or inner wires, one for
each of the commonly used RS-232 pins outlined in Figure 2-9.
The wires will be pinned "straight-through", in other
words, the wire from pin #2 on the DTE (PC) end will go "straight
through" to pin #2 on the DCE (modem) end, and so on with
the remaining pins.
Constructing modem cables or connecting any two devices for data
communications involves choosing between a variety of media types.
Coaxial cable, unshielded twisted pair, shielded twisted pair,
and fiber optic cable are but a few of the possible options. Although
media alternatives will be explored in detail in Chapter 6, it
is important to understand that media choices are present in every
data communications opportunity.
Many of the RS-232 pins and signals have a directionality to them.
In other words, either the terminal is informing the modem of
something by raising or lowering electrical voltages to a certain
pin, or the modem is informing the terminal of something by the
same means. Figure 2-9 outlines the directionality of the signals
of commonly used RS-232 pins in the columns labeled "FROM"
and "TO".
Figure 2-11 Parallel transmission physical interfaces
Given the transmission speed of today's modems, it is especially
important that PCs are equipped with the 16550 UART with its 16
byte buffer rather than previous UARTs which only contained a
1 byte buffer.
Figure 2-12 Breakout box
In general terms, based upon what we have learned thus far, we
could say:
It should be obvious that in order to better understand the processing
(P) which goes on in a modem, it is first necessary to
have a better understanding of (O), data transmission over
a "normal" or dial-up phone line.
As introduced in chapter one, a large switch in a telephone company
building called a central office or CO, connects your phone
equipment to the phone equipment of the party you wished to call
by finding an available circuit or path to your desired destination.
It is important to understand at this point that the CO switch
tries to find a path as quickly as possible to your destination.
The actual circuits it chooses represent the best path
available at that time. The particular circuits or path chosen
may vary from one occasion to another, even for calls to the same
location. Calls placed over dial-up lines through CO switches
which have connections built from available circuits are called
circuit-switched connections. In order to interface transparently
to the PSTN, modems must be able to dial and answer phone calls
to and from other modems.
The next important characteristic related to transmitting data
over a dial-up phone line has to do with how the data is represented
on that phone line. First, it is important to realize that today's
dial up phone network was originally designed to carry voice conversations
efficiently and with reasonable sound quality. This "efficiency
of design with reasonable sound quality" meant reproducing
a range of the frequencies of human speech and hearing just wide
enough to produce reasonable sound quality. That range of frequencies,
or bandwidth, is 3100 Hz (From 300 Hz to 3400 Hz), and
is the standard bandwidth of today's voice-grade dial up circuits
(phone lines). Hz is the abbreviation for hertz. One hertz is
one cycle per second. The higher the number of hertz or cycles
per second, the higher the frequency. Frequency, wavelength, hertz,
and cycles per second will be explained further later in this
chapter.
This 3100 Hz is all the bandwidth with which the modem operating
over a dial-up circuit has to work. Remember also that because
today's dial-up phone network was designed to be able to mimic
the constantly varying tones or frequencies which characterize
human speech, only these continuous, wave-like tones or frequencies
can travel over the dial-up phone network in this limited bandwidth.
The challenge for the modem, then, is to represent the discrete,
digitized 1's and 0's from the input (PC) side of the modem in
a continuous or analog form within a limited bandwidth
so that the data may be transmitted over the dial-up network.
Figure 2-13 summarizes the results of I-P-O analysis involving
modems and the PSTN.
Figure 2-13 I-P-O Analysis: Modems and the PSTN
Figure 2-14 Modulation vs. demodulation
This "normal" or "neutral" wave is called
a carrier wave as illustrated in Figure 2-15. The RS-232
pin #8 - Carrier Detect refers to this carrier or reference wave.
Modems generate carrier waves which are then altered (modulated)
to represent bits of data as 1's and 0's. When a local and remote
modem are trying to establish communications, you may have heard
a series of high-pitched screeches before the Carrier Detect indicator
lights on your modem. These screeches are the two modems trying
to detect a common carrier wave in order to establish communication.
Once the carrier wave has been detected by both modems, the actual
data transmission can begin as the modems manipulate this common
carrier wave to represent ones and zeroes.
Figure 2-15 Carrier Wave
How can the carrier wave be manipulated to represent ones and
zeroes? There are only three physical characteristics of this
wave which can be altered or modulated:
As will be seen later in this section, in some modulation schemes
more than one of these characteristics are altered simultaneously.
Each of the vertical lines in Figure 2-16 separate the one opportunity
to identify a 1 or 0 from another. These timed opportunities to
identify ones and zeroes by sampling the carrier wave are known
as signaling events. The proper name for one signaling event is
a baud.
Figure 2-16 Amplitude Modulation
The distance between the same spots on two subsequent waves is
called the wavelength. The longer the wavelength, the lower
the frequency and the shorter the wavelength, the greater the
frequency. Notice in Figure 2-17 how the higher frequency (shorter
wavelength) part of the wave represents a 1 and the lower frequency
(longer wavelength) part of the wave represents a 0, while amplitude
and phase remain constant. Again, the entire bit stream represents
the letter "A" in ASCII-7.
Figure 2-17 Frequency modulation
Figure 2-18 Phase modulation
Figure 2-19 Relationship between number of phase shifts
and number of potential detectable events
A simpler and perhaps clearer way to represent phase shifts as
illustrated in Figure 2-19 is through the use of constellation
points. Using a four quadrant representation of the 360 degrees
of possible phase shift, individual points represent each different
shifted wave. Note that when represented in a constellation diagram,
a phase shift of 270 degrees is represented as -90 degrees. Phase
shift modulation with four different phases is more properly referred
to as quadrature phase shift keying or QPSK.
However, in the second illustration in Figure 2-19, four detectable
events are possible for each signaling event, making it possible
to interpret two bits per baud. In this case, the bit rate or
transmission rate as measured in bps would be twice the baud rate.
More sophisticated modulation techniques are able to interpret
more than one bit per baud. In these cases, the bps is greater
than the baud rate. For example, is the baud rate of a modem was
2400 signaling events per second and the modem was able to interpret
2 bits per signaling event, then the transmission speed would
be 4800 bps. Mathematically, the relationship between baud rate
and transmission rate can be expressed as:
Transmission rate (bps) = Baud rate x bits/baud
By modifying a phase modulation technique such as that illustrated
in Figure 2-19, a modem can detect, interpret, and transmit more
than one bit per baud. The mathematical equation which describes
the relationship between the number of potential detectable events
and the numbers of bits per baud that can be interpreted is as follows:
Number of states = Number of potential detectable eventsbits/baud
To solve:
The answer is 2, meaning that 2 bits/baud can be interpreted at
a time. Two bits at a time are known as a dibit. By extending
the mathematical equation above, it should be obvious that:
How far can we go with increasing the number of phase shift angles
or potential detectable events ? One limiting factor to increasing
the bits/baud in phase shift modulation is:
Given that the modem is being used on a dial-up line of unpredictable
quality, how small (least number of degrees) a phase shift can
be reliably detected? Sixteen different phase shifts would require
reliable detection of phase shifts of as little as 22.5 degrees.
Remembering that phase is not the only wave characteristic which
can be varied, sixteen different detectable events can also be
produced by varying both phase and amplitude. Many of today's
high speed modems use a modulation technique which varies both
phase and amplitude known as quadrature amplitude modulation
or QAM
16QAM, with its sixteen different potential detectable
events, would allow 4 bits/baud or quadbits to be produced or
detected per signaling event. In this case the transmission rate
in bps would be 4 times the baud rate. Figure 2-20 illustrates
a representative set of constellation points and associated quadbits
for a QAM modulation scheme. Differences in phase are represented
in degrees around the center of the diagram while differences
in amplitude are represented by linear distance from the center
of the diagram. Each point is uniquely identified by combining
one of three potential amplitudes (.311V, .850V and 1.161V) with
one of twelve potential phase shifts (15, 45, 75, 105, 135, 175,
-165, -135, -105, -75, -45, -15 degrees). Obviously, all potential
combinations of these two sets of variables are not used in 16QAM.
Figure 2-20 QAM Constellation Points and quadbits
C = 2W
Considering a voice grade circuit: W = 3100 Hz, therefore
maximum theoretical data rate C = 6200 bps. However, this
fails to account for the ability of modern modems to interpret
more than one bit per baud by being able to distinguish between
more than just two symbols or potential detectable events. Taking
this ability into account, if we call the number of potential
detectable events (M), Nyquist's Theorem becomes:
C = 2W log2 M
Again, considering a voice grade circuit: if M = 16 potential
detectable events, then log2M = 4 and C = 24,800 bps.
(S/N)dB = 10 log (signal power/noise power)
S/N is expressed in decibels (dB). Decibels are a logarithmic
measurement using a reference of 0 dB for comparison. As a result,
a noise level of 10dB is 10 times more intense than a noise level
of 0 dB , a noise level of 20 dB is 100 times more intense than
a noise level of 0 dB , and a noise level of 30 dB is 1000 times
more intense than a noise level of 0 dB .
Shannon found that the higher the data rate, the more interference
is caused by a given amount of noise, thus causing a higher error
rate. This should make sense since at higher data rates, more
bits are traveling over a circuit in a fixed length of time, and
a burst of noise for the same length of time will effect more
bits at higher data rates. By taking into account the signal to
noise ratio, Shannon expresses the maximum data rate of a circuit
(C) as:
C = W log2 (1 + S/N)
Considering a voice grade circuit: if W = 3100 Hz and the
S/N = 30dB , and remembering that a S/N of 30dB
yields a ratio of 1000, then C = 30,894 bps. Depending
on the value inserted into Shannon's Law for S/N or for
W, C will vary accordingly. 24,000 bps is also a
common value for C using Shannon's Law. By substituting
different values for W and for S/N , it should become
obvious that data channel capacity (C) can be more drastically
effected by changes in bandwidth (W) , than by changes
in the signal to noise ratio (S/N).
Finally, it should be noted that there are numerous other line
impairments such as attenuation, delay distortion, and impulse
noise which Shannon's Law does not take into account. As a result,
the data channel capacity (C) derived from Shannon's Law
is theoretical and is sometimes referred to as error-free capacity.
Obviously, the modems must have a reliable way to know exactly
when to sample the line for data. Somehow, the local and remote
modems must establish and maintain some type of timing between
them so that these detectable events are produced, transmitted,
and detected accurately. There are two main alternatives to establishing
and maintaining the timing for the sampling of detectable events.
These two timing alternatives are known as:
Figure 2-21 summarizes some important characteristics about asynchronous
and synchronous transmission methods. The most noticeable difference
is that the synchronization, or detectable event timing, itself
is re-established with the transmission of each character in asynchronous
transmission via the use of start and stop bits and with each
block of characters in synchronous transmission. In asynchronous
transmission, there may be 1, 1.5, or 2 stop bits. The timing
in the case of synchronous transmission is supplied by a clock
which may be supplied by either the remote or local modem, or
by the carrier.
Secondly, when comparing idle time activity, synchronization is
maintained, thanks to the ever-present clocking signal, in synchronous
transmission and dropped in asynchronous transmission while no
characters are being transmitted. The effect of these characteristics
on transmission efficiency is illustrated in Figure 2-21
Figure 2-21 Asynchronous vs. Synchronous transmission
In terms of application, PC modems use asynchronous transmission
as do asynchronous terminals such as a VT-100 to an asynchronous
minicomputer such as a VAX. Synchronous transmission is used between
mainframe terminals such as IBM 3270 and IBM mainframe computers.
Synchronous transmission is also used for most high-speed WAN
services such as 64Kbps or T-1.
A four-wire circuit is comprised of two wires capable of
simultaneously carrying a data signal each with its own dedicated
ground wire. Typically, four wire circuits are reserved for leased
lines, otherwise known as dedicated lines or private lines.
These circuits bypass telephone company switching equipment. They
have no dial tone and are always operational between the locations
specified by the customers ordering the leased lines.
Modems which interfaced to a dial-up circuit had to support this
half-duplex transmission method. What this meant was that once
the two modems completed initial handshaking, one modem
would agree to transmit while the other received. In order for
the modems to reverse roles, the initially transmitting DTE (terminal
or computer) drops its RTS (request to send) RS-232 Pin # 4 ,
the transmitting DCE (modem) drops its CTS (clear to send) RS-232
Pin # 5, and perhaps its carrier wave. Next, the initially receiving
DTE must raise RTS, the initially receiving DCE (modem) must generate
a carrier wave and raise CTS, and the role reversal is complete.
This role reversal is known as turnaround time and can
take two-tenths of a second or longer. This may not seem like
a very long time, but if this role-reversal needed to be done
several thousand times over a long distance circuit, charged by
usage time, it may have a large dollar impact.
Full-duplex transmission supports simultaneous data signaling
in both directions. Full-duplex transmission might seem to be
impossible on two-wire circuits. Until the advent of the V.32
9600 bps full duplex modem, the only way to get full duplex transmission
was to lease a four wire circuit. Two wires (signal & ground)
were for transmitting data and two wires (signal & ground)
were for receiving data. There was no "role reversal"
necessary and therefore, no modem turnaround time delays.
Modems manufactured to the CCITT's V.32 standard (and the later
V.34 standard) can transmit in full-duplex mode, thereby
receiving and transmitting simultaneously over dial-up two-wire
circuits. These modems use sophisticated echo cancellation
techniques and, at least when they were first introduced,
were significantly more expensive than slower, half-duplex modems.
Figure 2-22 highlights the differences between half-duplex and
full-duplex transmission.
Figure 2-22 Half-Duplex vs. full-duplex transmission
An analog or voice grade leased line is a normal voice
grade line which bypasses the carrier's switching equipment. When
the service is ordered, you must name the two or more locations
which are to be connected. A new circuit is installed into these
locations with a new RJ-11 jack interface added at each circuit
location. There is no dial-tone on this new line because it does
not go through any switching equipment. With only two locations
connected, the circuit is called point-to-point, with more
than two locations, the circuit is called multipoint.
For leased lines, charges include both an installation charge
as well as a flat monthly charge. The installation charge may
be significant, especially if the circuit crosses LATAs. The flat
monthly fee does not vary by usage, unlike the dial-up voice grade
circuits. The circuit is exclusively yours to use, 24 hours per
day, 7 days per week.
When ordering a narrowband digital circuit the speed of that circuit
must be specified to the phone company at the time the order is
placed. Typical choices are:
This "declaring your speed" brings up a key difference
in operation of analog and digital circuits.
When an analog circuit, dial-up or leased, degrades or has some
kind of transmission impairment, many modems use fallback
or lower speeds automatically and continue with data transmissions.
With a digital circuit, the speed of the circuit is set at the
CO, the CSU/DSUs are set to the speed of the circuit, and if there
is a problem with the line, the data transmission ceases. It should
be pointed out that digital circuits tend to be more error free
than analog circuits. Simply stated, the key determining factor
in the transmission speed of an analog circuit is the modem, while
the key determining factor in the case of a digital circuit is
the configured speed of the circuit itself.
ISDN service, Switched 56K service, or any of the digital services
listed in Figure 2-23, are not necessarily available at any given
home or business location. One should never assume that any particular
transmission service is available at any network location when
performing network planning and design.
| |||||||||||||||||||||||||||||||||||
|
Wiley MIS Page | Wiley College Page | Wall Street Journal Business Extra Copyright © 1997 Wiley & Sons. All rights reserved. |
