Xem mẫu
- Networks and Telecommunications: Design and Operation, Second Edition.
Martin P. Clark
Copyright © 1991, 1997 John Wiley & Sons Ltd
ISBNs: 0-471-97346-7 (Hardback); 0-470-84158-3 (Electronic)
Data Network Principles
and Protocols
We have described the binary form in which data are held by computer systems, and how such
data are conveyed over digital line transmission systems, but we shall need to know more than
this before we can design the sort ofdevices which can communicatesensibly with one another in
somethingequivalent to human conversation.Inthischapter we shalldiscuss indetailthe
conveyance of data between computer systems, the networks
required, the
and so-called
‘protocols’ they will need to ensure that they are communicating properly. In the second half of
the chapter some practical data and computer network topologies and the equipment needed to
support them are described.
9.1 COMPUTER NETWORKS
Between 1950 and 1970 computers were large and unwieldy, severely limited in their
power and capabilities, and rather unreliable. Only the larger companies could afford
them, and they were used for batch-processing scientific, business or financial data on a
largescale. Data storageinthosedayswaslaborious,limitedincapacity,longin
preparation, not at all easy to manage. Many storage mechanisms (e.g. paper tape and
punched cards) were very labour-intensive; they were difficult to store, and prone to
damage. Even magnetic tape, when it appeared, had its drawbacks; digging out some
trivial information ‘buried’ in the middle of a long tape was a tedious business, and the
tapes themselves had to be painstakingly protected against data corruption and loss
caused by mechanical damage or nearby electrical and magnetic fields.
Computing was for specialists. Computer centre staff lookedafterthe hardware,
while software experts spent hours
long improving computer
their programmes,
squeezing every last drop of ‘power’ out of the computers’ relatively restricted capacity.
By the mid-1970s all this began to change, and very rapidly. Cheap semiconductors
heralded the appearance of the microcomputer, which when packaged with the newly
developed floppy diskette systems, opened a new era of cheap and widespread com-
puting activity. Personal computers ( P G ) began to appear on almost every manager’s
177
- 178 PRINCIPLES
NETWORK DATA AND PROTOCOLS
desk, and manyeven invaded peoples’ homes. Suddenly, computing waswithin reach of
the masses, and the creation and storage of computer data was easy, cheap and fast.
All sorts of individuals began to prepare their own isolated databases and to write
computer programmes for small scale applications, but these individuals soon recog-
nized the need to share information and to pass data between different computers. This
could be done by transferring floppy diskettes from one machineto another, but as time
went on that method on its own proved inadequate. Therewas a growing demand for
more geographically widespread, rapid, voluminous data transfer. More recently, the
demands of distributed processing computer networks comprising clients and servers
have created a boom in demand for data networks.
A very simple computer or data network consists of a computer linked to a piece of
peripheral equipment, such as a printer. The link is necessary so that the data in the
computer’s memory can be reproduced on paper. The problem is that a ‘wires-only’
direct connection of this nature is only suitable for very short connections, typically up
.to about 20 metres. Beyond this range, some sort of line driver telecommunications
technique must be used. A number of techniques are discussed here. A long distance
point-to-pointconnectionmay be made using modems. A slightly more complex
computer network might connect a number of computer terminals in outlying buildings
back to a host (mainframe computer) in a specialized data centre. Another network
might be a Local Area Network (or L A N ) , used in an office to interconnect a number of
desktop computing devices, laser printers, data storage devices (e.g. file servers), etc.
More complex computer networks might interconnect a number of large mainframe
computers in the major financial centres of the world, and provide dealers with ‘up-to-
the-minute’ market information.
The basic principlesof transmission, asset out in the early chapters this book, apply
of
So
equally to datawhich are communicated around computer networks. circuit-switched
networks or simple point-to-point lines may also beused for data communication. Data
communication, however, makes more demandson its underlying network than a voice
or analogue signal service, and additional measures are needed for coding the data in
preparation for transmission, and in controlling the flow of data during transmission.
Computers do not have the same inherent ‘discipline’ to prevent them talking two at a
time. For this reason special protocols are used in data communication to make quite
sure that information passing between computers is correct,complete and properly
understood.
9.2 BASIC DATA CONVEYANCE: INTRODUCING THE DTE
ANDTHE DCE
As we learned in Chapter 4, data are normally held in a computer or computer storage
medium in a binary code format, as a string of digits with either value ‘0’ or value ‘l’.
A series of suchbinarydigitscanbe used to representalphanumericcharacters
(e.g. ASCII code), or graphical images, such as those transmitted facsimile machines,
by
video or multimedia signals.
In Chapter 5 we went on to discuss the principles of digital transmission, and found
that it was ideal for the conveyance of binary data. Digitaltransmission has become the
- CONVEYANCE:
DATA
BASIC INTRODUCING THE DTE AND THE DCE 179
backbone of both private and public networks. However, despite increasing
the
availability and ideal suitability of digital transmission for data communication, it is
unfortunately not always available. In circumstances where it is not, digitally-oriented
computer information must instead be translated into a form suitable for transmission
across an analogue network. This translation is carried out by a piece of equipment
called a modulatorldemodulator, or modem forshort.Modems transmit data by
imposing the binary (or digital) data stream onto an audio frequency carrier signal.The
process is very similar to that used in thefrequency division multiplexing of voice
channels described in Chapter 3.
Figure 9.1 illustrates two possible configurations for data communication between
two computers using either a digital or an analogue transmission link. The configura-
tionslook very similar,comprising computers
the themselves (these are specific
examples o data terminal equipment ( D T E ) ) ;sandwiched between them in each caseis a
f
line and a pair of data circuit terminating equipments (DCE).
A digital DCE (Figure9.l(a)) connectsthecustomer'sdigital DTE to adigital
transmission line, perhaps provided by the public telecommunications operator (PTO).
The DCE provides several network functions. In the transmit direction, it regenerates
the digital signal provided by the DTE and converts it into a standardized format,level
and line code suitable for transmission on the digital line. More complex DCEs may
also interpret the signals sent by the DTE to the network to indicate the address desired
DT E DTE
2
'
TCE/
Digital
line
DCE 2
Computer
,
Digital
information
(a) D i g i t a l Line connection
D i g i t ai ln f o r m a t i o n
DTE DTE
DCE modula
ted DCE
analoguesignal
Computer Modem < Modem I
Computer
-
2 way analogue
transmission
link
Figure 9.1 Point-to-point connection of computers. DTE, Data Terminal Equipment; DCE,
Data Circuit Terminating Equipment
- 180 PRINCIPLES
NETWORK DATA AND PROTOCOLS
when establishing a connection, and help in setting up the call. In the receive direction,
the DCE establishes a reference voltage for use of the DTE and reconverts the received
line signal into a form suitable forpassing to the DTE. In addition, it uses the clocking
signal (i.e. the exact bit rate of the received signal) as the basis for its transmitting bit
rate.The received clockingsignal is usedbecausethis is derivedfromthe highly
accurate master clock in the PTO’s network. Two interfaces need to be standardized.
These are the DTE/DCE interface and the DCE/DCE interface.
DigitalDCEscan be used toprovidevariousdigitalbitspeeds,fromas low as
2.4 kbit/s, through the standard channel of 64 kbit/s, right up to higher order systems
such as 1.544Mbit/s (called T1 or DSl), 2.048 Mbit/s (called El), 45 Mbit/s (called T3
or DS3) or higher. When bit speeds below the basic channel bit rate of 64 kbit/s are
required by the DTE, then the DCE has an additional function to carry out, breaking
down the line bandwidth of 64 kbit/s into smaller units in a process called sub-rate
multiplexing. The process derives a numberof lower bit rate channels, such as kbit/s, 2.4
4.8 kbit/s, 9.6 kbit/s, etc., some or all of which may be used by a number of different
DTEs.
In their various guises, digital DCEs go by different names. In the United Kingdom,
BritishTelecom’s Kilostream digitalprivateline service uses a DCE called a NTU
(networkterminatingunit). Meanwhile,intheUnitedStates,AT&T’s digitaldate
system ( D D S ) and similar digital line services are provided by means of channel service
units ( C S U ) or data service units ( D S U ) . Sometimes the term LTU (line terminating
unit) is used.
In the caseof analogue transmission lines, the DCE (Figure 9.1 (b)) cannot on the
rely
network to provide accurate clocking information, because the bit rate of the received
signal is not accurately set by the PTO’s analogue network. For this reason, an internal
clock is needed within the DCE in order to maintain an accurate transmitting bit rate.
The other major difference between analogue and digital DCEs is that analogue DCEs
(modems) need to convert and reconvert digital signals received from the DTE into an
analogue signal suitable for line or radio transmission.
9.3 MODULATION OF DIGITAL INFORMATION OVER ANALOGUE
LINESUSINGAMODEM
Three basic datamodulation techniques are used in modems for conveying digital end
user informationacrossDCE/DCE interfaces
comprising analoguelines or radio
systems, buttherearemoresophisticated versions of eachtype and even hybrid
versions,combiningthevarioustechniques.Each modem is inrealityaspecialized
device for a specific line or radio system type. The three techniques are illustrated in
Figure 9.2 and described below.
Amplitude modulation
Modems employing amplitudemodulation altertheamplitude of thecarriersignal
between a set valueand zero (effectively ‘on’ and ‘off ’) according to the respective value
‘1’ or ‘0’ of the modulating bit stream. Alternatively two different, non-zero values of
amplitude may be used to represent ‘1’ and ‘0’.
- MS HIGH BIT RATE 181
Phase Phase
180’ change 0’ change
[ a ) Amplitude
modulation ( C ) Phase
modulation
l b) Frequency modulation
Figure 9.2 Datamodulationtechniques
Frequency modulation
Infrequency modulation (Figure 9.2(b)), it is the frequency of the carrier signal that is
altered to reflect the value ‘1’ or ‘0’ of the modulating bit stream. The amplitude and
phase of the carrier signal is otherwise unaffected by the modulation process. If the
number of bits transmitted per second is low, then the signal emitted by a frequency
modulated modem is heard as a ‘warbling’ sound, alternating between two frequencies
of tone. Modems usingfrequencymodulation are more commonly called FSK, or
frequency shift key modems.A commonform of FSK modem uses four different
frequencies(ortones),two for the transmit direction and two for the receive. This
allows simultaneous sending and receiving (i.e. duplex transmission)of data by the
modem, using only a two-wire circuit.
Phase modulation
In phase modulation (Figure 9.2(c)), the carrier signal is advanced or retarded in its
phase cycle by the modulating bit stream. At the beginning of each new bit, the signal
will either retain its phase or change its phase. Thus in the example shown the initial
signal phase represents thevalue ‘l’. The change of phase by 180” represents next bit ‘0’.
In the third bit period the phase does not change, so the value transmitted is ‘l’. Phase
modulation (often called phase sh$t keying or P S K ) is conducted by comparing the
signal phase in one time period to that in the previous period; thus it is not the absolute
value of the signal phase that is important, rather the phase change that occurs at the
beginning of each time period.
9.4 HIGH BIT RATE MODEMS
The transmission of high bit rates can be
achieved by modems in one of two ways. One is
t o modulate the carrier signal at a rate equal to the high bit rate of change of the
- 182 PRINCIPLES
NETWORK DATA AND PROTOCOLS
I I Bit rate
( 2 bl t per second1
Bit stream
1
Pair of b i t s Iv
10 10
v
11 00
v
10 01
I '
Modemllne
signal f , (10)
(frequency Baud r a t e
and 2 - b i t fz(01)
value) I lsec I 00 ( lsecond per
change )
Figure 9.3 Multilevel transmission and lower baud rate
modulating signal. Now the rate (or frequency) which we fluctuate the carriersignal is
at
called the baud rate, and the disadvantage of this first method of high bit rate carriage is
the equally high baud rate that it requires. The difficulty lies in designing a modem
capable of responding to the line signal changes fast enough. Fortunately an alternative
method is available inwhich the baud rate lower than thebit rate of the modulatingbit
is
stream. The lower baud rate is achieved by encoding a number of consecutive bits from
the modulating stream to be represented by a single line signal state. The method is
called multilevel transmission, and is most easily explained using a diagram. Figure 9.3
illustrates a bit stream of 2 bits per second(2 bit/s being carried by a modem which uses
four different line signal states. The modem is able to carry the bit stream at a baud rate
of only 1 per second (1 Baud)).
The modem used in Figure 9.3 achieves a lower baud rate than the bit rate of the data
transmitted by using each of the line signal frequencies f 1, f2, f 3 and f4 to represent
two consecutive bits rather than just one. This means that the line signal always has to
be slightly in delay over the actual signal (by at least one bit as shown), but the benefit is
that the receiving modem will have twice as muchtime to detect and interpret the
datastream represented by the received frequencies. Multi-level transmission is invari-
ably used in the design of very high bit rate modems.
9.5 MODEM 'CONSTELLATIONS'
At this point we introduce the concept of modem constellation diagrams, because these
assistin theexplanation of morecomplex amplitude and phase-shift-keyed ( P S K )
modems. Figure 9.4 illustrates a modem constellation diagram composed of four dots.
Each dot on the diagram represents the relative phase and amplitude of one of the pos-
sible line signal states generated by the modem. The distance of the dot from the origin
- MODEM 'CONSTELLATIONS' 183
Signal amplitude
0 90 180 270 360
( a ) L - signal
state
constellation ( b ) Slgnal
phase,commencing at 0
'
Signal
phase
( c ) Signal phase,cammencing at 90' ( d ) The four signals
line
Figure 9.4 Modem constellation diagram
of the diagram axes represents the amplitude of the signal, and the angle subtended
between the X-axis and a line from the point of origin represents the signal phase
relative to the signal state in the preceding instant of time.
Figure 9.4(b) and (c) together illustrate what we mean by signal phase. Figure 9.4(b)
showsasignalof 0" phase, in which thetimeperiod starts withthe signal at zero
amplitude and increases to maximum amplitude. Figure 9.4(c), by contrast, shows a
signal of 90" phase, which commences further on in the cycle (in fact, at the 90" phase
angle of the cycle). The signal starts at maximum amplitude but otherwise follows a
similar pattern. Signal phases, for any phase angle between 0 and 360" could similarly
be drawn. Returning to the signals represented by the constellation of Figure 9.4(a)we
can now draweach of them, as shownin Figure 9.4(d) (assuming that each of them was
preceded in the previous time instant by a signal of 0 phase). The phase angles in this
case are 45", 135", 225", 315".
Wearenowreadyto discusscomplicated
a butcommonmodemmodulation
technique known as quadrature amplitude modulation, or Q A M . Q A M is a technique
usingacomplexhybrid of phase (or quadrature) as well as amplitudemodulation,
hence the name. Figure 9.5 shows a simple eight-state form ofQAM in which each line
signal state representsathree-bitsignal(values noughtto seven inbinarycan be
representedwithonlythreebits). The eight signal states are a combination of four
different relative phases and two different amplitude levels. The table of Figure 9.5(a)
relates the individual three-bit patterns to the particular phases and amplitudes of the
signals that represent them. Note: Figure 9.5(b) illustrates the actual line signal pattern
that would result if we sent the signals in the table consecutively as shown. Each signal
- 184 PRINCIPLES
DATA NETWORK AND PROTOCOLS
l-
iI
Line signal
Bit
omblnation implitude Phase
shift
000 low 0
' l
00 1 high 0
'
01 0 Iow 90'
01 1 high 90'
100 I ow 180'
101 high 180' 000.010 011. 111 .l00 001 110 101 . l 0 1
'
110 low 270'
(b) Typical line signal
11 1 high 270'
( a ) B i t combinations and
line signal attributes
High
Amplitude
180' X
27'
0
t
( c ) Modem constellation
Figure 9.5 Quadrature amplitudemodulation
is shown in the correct phase state relative to the signal in the previous time interval
(unlike Figure 9.4 where we assumed that each signal individuallyhad been preceded by
one of 0 phase). Thus in Figure 9.5(b) the same signal state is not used consistently to
convey the same three-bit pattern, because the phase difference with the previous time
period is what counts, not the absolute signal phase. Hence the eighth and ninth time
periods in Figure 9.5(b) both represent the pattern 'lOl', but a different line is used,
180" phase shifted.
Finally, Figure 9.5(c) shows the constellation of this particular modem.
To finish off thesubject of modemconstellations,Figure9.6presents,without
discussion, the constellation patternsof a coupleof very sophisticated modems,specified
by ITU-T recommendationsV.22 bis and V.32 as a DCE/DCE interface. As in Figure9.5,
the constellation pattern would allow the interested reader to work out the respective 16
and 32 line signal states. Finally, Table 9.1 lists some of thecommon modem types and
their uses. When reading the table, bear in mind that synchronous and asynchronous
operation is to be discussed later in chapter, and that
the half-duplex means that two-way
transmission is possible but in only one direction at a time. This differs from simplex
operation (as discussed in Chapter 1) where one-way transmission only is possible.
- MODEM ‘CONSTELLATIONS’ 185
( a ) V 2 2 bis ( 3 a m p l i t u d(e s),
b V 3 2 15 amplitudes,
phases) 28 12 p h a s e s )
Figure 9.6 More modemconstellations
Table 9.1 Commonmodemtypes
Modem type Synchronous (S)
(ITU-T Modulation speed
Bit or asynchronous Full or half
Circuit
type
recommendation) type (bit/s) (A) operation required
duplex
~ ~~~~~~
v.21 FSK up to 300 A Full 2w telephone line
v.22 PSK 1200 S or A Full 2w telephone line
V.22 bis QAM 2 400 S or A Full 2w telephone line
V.23 FSK 60011 200 A Full 4w leaseline
Half 2w telephone line
V.26 2 400 S Full 4w leaseline
V.26 bis 2 400/ 1 200 S Half 2w telephone line
V.27 4 800 S Half 2w leaseline
V.21 ter 4 800 S Half 2w leaseline
V.29 9 600 S Full 4w leaseline
V.32 up to 9 600 S or A Full 2w telephone line
v.33 14 400 S or A Full 2w telephone line
v.34 up to 33600 S or A Full 2w telephone line
v.35 48 000 wideband Full groupband
leaseline
V. 36 AM 48 000 wideband Full groupband
leaseline
v.37 AM 12000 wideband Full groupband
leaseline
V.42 convert Error
synchronous correcting
to protocol
asynchronous
format
V.42 bis up to 30 kbit/s in - Data
association with compression
V.32 modem technique
- 186 PRINCIPLES
NETWORK DATA AND PROTOCOLS
9.6 COMPUTER-TO-NETWORK INTERFACES
Returning nowto Figure 9. I , we can see that we have discussed in some detail the method
of data conveyance overthe telecommunications line between one DCE and the other (or
between one modem and the other) but what about the that connects computer
interface a
to a DCE or modem? This interface is of a generic type (DTE/DCE, i.e. between DTE and
DCE) and can conform to any one of a number of standards, as specified by various
organizations. The interface controls the flow of data between DTE and DCE, making
sure that the DCE sufficient instructions to
has deliver the data correctly. also allows the
It
DCE to prepare the distant DTE confirm receipt of data if necessary. Physically, the
and
interface usually takes the form of a multiple pin connection (plug and socket)typically
with 9,15,25 or pins arranged in a ‘D-formation’
37 (so-called D-socket or sub-D-socket).
Computer users will be familar with the type sockets shown in Figure 9.7.
of
The interface itself is designed either for parallel data transmission or serial data
transmission, the latter being the more common. The two different methods of trans-
mission differ in the way each eight-bit data pattern is conveyed. Internally, computers
operate using the parallel transmisssion method, employing eight parallel circuits to
carry one bitof information each. Thus during one time intervalall eight bits of the data
pattern are conveyed. The advantage of this method is the increased computer process-
ing speed which is made possible. The disadvantage is that it requires eight circuits
insteadofone.Paralleldatatransmission is illustratedinFigure9.8(a),wherethe
pattern ‘10101110’ is being conveyed over the eight parallel wires of a computer’s data
bus. Note: nowadays 16-bit, and even 32-bit data buses are used in the most advanced
computers. Examples of parallel interfaces are those specified by ITU-T recommenda-
tions V.19 and V.20.
Serial transmission requires only one transmission circuit, and so is far more cost-
effective for data transmisssion on long links between computers. The parallel data on
the computer’s bus is converted into a serial format simply by ‘reading’ each line of the
bus in turn. The same pattern ‘101011 10’is being transmitted in a serial manner in
Figure 9.8(b). Note that the baud rate needed for serial transmission is much higher
than for the equivalent parallel transmission interface.
Many DTE-to-DCE (or even short direct DTE-to-DTE connections) use a serial
transmission interface conforming to oneof ITU-T’s suites of recommendations, either
X.21 or X.2lbis. The basic functions of all types of DTE/DCE interface are similar:
0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 ?
?
25-pin
(female)
D-socket (IS0 21 10) 15-pin (female)
sub-D-socket (IS0 4903)
used on DCE for DCE/DTE used onDCE for DCE/DTE
connection (RS232 or V.24N.28) connection (X.21, V.10 or V . l l )
Figure 9.7 Plug and socket arrangement for common DTE/DCE cables
- COMPUTER-TO-NETWORK INTERFACES 187
0 clocking and synchronizing the data transfer
0 regulating the bit rate, so that the receiving device does not become swamped with
data.
In Figure 9.9 we summarize the complete suite ITU-T recommendationsdetailing the
of
most important physical layer DTE/DCE interfaces. These are those defined by ITU-T
recommendations X.21 and X.21 bis as shown.
Circuit
voltage
1
f
n 1
Computer
Computer
(integrated
’ ‘data bus’
Circuit)
t
Circuit
number
(a] P a r a l l er a n s m i s s i o n
tl
Direction o f transmission
0 1 1 1 0 1 0 1
DTE
m Line
( b )e r i a r a n s m i s s i o n
S tl
Figure 9.8 ‘Parallel’ and ‘serial’ transmission
- 188 PRINCIPLES
DATA NETWORK AND PROTOCOLS
/&,
procedure
RS232
functional interface
l I
electrical interfa;ce
Fm
f 3 b f15V
I I
25-ph &i
Pn
Figure 9.9 Common interfaces for DTE/DCE attachment
The ITU-T recommendations listed in Figure 9.9 define for each of the interfaces:
themechanicalnature of theplugs andsockets (i.e. theirphysicalstructure),the
electricalcurrentsandvoltagesto be used,thefunctionalpurpose or meaning of
on
the various circuits which appear the various pinsof the plug/socket connection, and
the valid pin combinations and order (or procedure) in which the various functions
may be used.
There are two broad groups physical layer interfaces. The first broad group are the
of
group sometimesloosely referred to as V.24 (RS-232) or X.21bis which were developed
in conjunction with modems and are relativelylow speed devices (up to about 28 kbit/s
and about 15 m cable length) allowing computer data tobe transferred to modems for
carriage across analogue telephonelines. The second group is the X.21 suite (sometimes
only the subsets X.24 or V.ll are used). This second group was developed later for
higher speed lines, which became possible in the advent of digital telephone leaslines
(64 kbit/s and above and up to 100 m cable length).
In Europe, the standard interfaces offered by modems and digital leaseline DCEs
(variously called line terminating unit ( L T U ) ,network terminating unit ( N T U ) , channel
service unit (CSU), DSU(data service unit)) are V.24 for analogue leaselines and X.21
(or the subsets X.24 V. 11) for digital lines. In North AmericaRS-232 (V.24) is widely
or
used for analogue lines, but V.35 and V.36 predominate over X.21 for digital lines.
- SYNCHRONIZATION 189
You might wonder how you connect a V.35 DTE in the USA to an X.21 DTE in
Europe. answer
The is quitestraightforward. You order an international digital
leaseline with V.35 DCE interface in the USA and X.21 interface in Europe. The fact
that the two DTE-DCE connections have different forms is immaterial.
9.7
SYNCHRONIZATION
An important component of any data communications system is the clocking device.
The data consists of a square, ‘tooth-like’ signal, continuously changing between state
‘0’ and state ‘l’, as we recall in Figure 9.10.
The successful transmission of data depends not only on the accurate coding of the
transmitted signal (Chapter4), but also on the ability of the receiving device to decode
the signal correctly. This calls for accurate knowledge (or synchronization) of where
each bit and each message begins and ends.
The receiver usually samples the communication line at a rate much faster than that
of the incoming data, thus ensuring a rapid response to any change in line signal state,
as Figure 9. 1 shows. Theoretically itis only necessary to sample the incomingdata at a
l
rate equal to the nominal bit rate, but this runs a risk of data corruption. If we choose
to sample near the beginning or end of each bit we might lose or duplicate data as
Figures 9.1 l(a) and (b) show, simply as the result of a slight fluctuation in the time
duration of individual bits. Much faster sampling ensures rapid detectionof the start of
each ‘0’ to ‘1’ or ‘1’ to ‘0’ transition, as Figure 9. l l(c) shows. The signal transitions are
interpreted as bits.
Variations in the time duration of individual bits come about because signals are
liable to encounter time shifts during transmission, which mayor may not be the same
for all bits within the total message. These variations are usually random and they
combine to create an effect known as jitter, which can lead to incorrect decoding of
incoming signals when the receiver sampling rate is too low, as Figure 9.12 shows.
Jitter, together with a slight difference in the timing of the incoming signal and the
receiver’s low sampling rate have created an error in the received signal, inserting an
State ‘1’
State ‘0’
Figure 9.10 Typical data signal
t t t t tttttttt
Lsat)m p l e
(aa e ( b l Early
sample ( c ) High sample
rate
Figure 9.11 Effect of sample rate
- 190 PRINCIPLES
NETWORK DATA AND PROTOCOLS
Actualdatapattern :
1 0 1 0 1 0 1 0 1 0
I l r I I I l
I I I I I I I l I
I I I I I I I I I Signal
t f f f f f t t t t t Sampling
instant
Interpreted recelved signal :
1 0 0 1 0 1 0 1 0
J-2
Error, caused
by ‘ j i t t e r ’
Figure 9.12 The effect of jitter
extra ‘l’. Thus the signal at the top is intended to represent only a ten bit pattern, but
because the bit durations are not exactly correct, (due to jitter), the pattern has been
misinterpreted. An extra bit has been inserted by the receiving device.
To prevent an accumulation of errors over a period of time, we use a sampling rate
higher than the nominal data transfer rate as already discussed; we also need to carry
out a periodic synchronization of the transmitting and receiving end equipments.
The purpose of synchronization is to remove all short, medium and long term time
effects. In the very short term, synchronization between transmitter and receiver takes
place at a bit level, by bit synchronization, which keeps the transmitting and receiving
clocks in step, so that bits start and stop at theexpected moments. In the medium term
it is also necessary to ensure character or word synchronization, which prevents any
confusion between the lastfew bits of one character and the first few bits of the next. If
we interpret the bits wrongly, we end up with the wrong characters. Finally there is
frame synchronization, which ensures data reliability (or integrity) over even longer time
periods.
9.8 BIT SYNCHRONIZATION
Figure 9.13 shows two different pulse transmission schemes used for bit synchroniza-
tion. The first, Figure 9.13(a), is a non-return to zero ( N R Z ) code which looks like a
Jm
- , m (a)
Non-return to
--
zero ( N R 2 1 code
1 1 0 1 0 0 1 0
1
+
Return to zero
(b)
( R Z ) code
1 1 0 1 0 0 1 0
Figure 9.13 Methods of bit synchronization
- CHARACTER
SYNCHRONIZATION:
SYNCHRONOUS AND ASYNCHRONOUS
DATA
TRANSFER 191
string of ‘1’ and ‘0’ pulses, sent in the manner with which we are already familiar.
In contrast Figure9.13(b) is a return-to-zero ( R Z )code, in which ‘l’s are represented by
a short pulse which returns to zero at the midpoint. RZ code therefore provides extra0
to 1 and 1 to 0 transitions, most noticeably within a string of consecutive ‘l’s. By so
doing,itbettermaintainssynchronizationof clock speeds (or bitsynchronization)
between the transmitter and the receiver. A number of alternative techniques also exist.
The technique used in a given instance will depend on the design of the equipment and
the accuracy of bit synchronization required. More often than not it is the interface
standard that dictateswhich type will be used. The synchronization codeused widely on
IBM SDLC networks is called NRZI (non-return to zero inverted).These and a number
of other line codes are illustrated in Chapter 5.
9.9 CHARACTER SYNCHRONIZATION: SYNCHRONOUS
AND
ASYNCHRONOUS DATATRANSFER
Data conveyance a overtransmission link can be either synchronous (in which
individual data characters (or at least a predetermined number of bits) are transmitted
at a regular periodic rate) or asynchronous mode (in which the spacing between the
characters or parts of a message need not be regular). In asynchronous data transfer
each datacharacter(represented say by an eight-bit ‘byte’) is preceded by a few
additional bits, which are sent to mark (ordelineate) the start of the eight-bit string to
the receiving end. This assures that character synchronization of the transmitting and
receiving devices is maintained.
Between characterson an asynchronous transmissionsystemthe line is left ina
quiescent state, and the system is programmed to send a series of ‘l’s during this period
to ‘exercise’ the line and not to generate spurious start (value ‘0’). When a character
bits
(consistingofeightbits) is ready to besent,thetransmitterprecedestheeightbit
pattern with an extra start bit (value ‘O’), then it sends the eight bits, and finally it
suffixes the pattern with two ‘stop bits’. both set to ‘l’. The total pattern appears as in
Figure 9.14, where the user’s eight bit pattern 11010010isbeing sent. (Note: usually
nowadays, only one stop bit used. This reduces the overall number of bits
is which need
to be sent to line to convey the same information by 9%.)
In asynchronous transmission, the line is not usually in constant use. The idle period
between character patterns (thequiescent period) is filled by a string of 1S. The receiver
can recognize the start of each character when sent by the presence of the start bit
transition(fromstate ‘1’ tostate ‘0’). The followingeightbitsthenrepresentthe
character pattern.
The advantage of asynchronous transmission lies in its simplicity. The start and stop
bits sent between characters help to maintain synchronization without requiring very
accurate clocking, so that devices can be quite simple andcheap.Asynchronous
transmission is quite widely between
used computer terminals andthecomputers
themselves because of the simplicity of terminal design and its consequent cheapness.
Given that human operators type at indeterminate speeds and sometimes leave long
pauses between characters,it is ideally suited to this use. The disadvantage of
asynchronous transmission lies in its relatively inefficient use of the available bit speed.
- 192 PRINCIPLES
DATA NETWORK AND PROTOCOLS
Start stop
bit bits
State‘O‘
S t a t e ‘1’
Quiescent
period 1 1
Figure 9.14
8- bit pattern
0 1 0 0 1
4 0
Asynchronous data transfer
Quiescentperiod,
or next
character
SY N User d a t a ( m a n y bytes) SYN SYN
Figure 9.15 Synchronous data transfer
As we can see fromFigure 9.14, out of eleven bitssentalongthe line, only eight
represent useful information.
In synchronous data transfer, the data areclocked at a steady rate. A highly accurate
clock is used at both ends, and a separate circuit may be used to transmit the timing
between the two. Provided thatall the data bit patterns are of an equal length, the start
ofeach is known to followimmediatelythepreviouscharacter. Theadvantage of
synchronous transmission is that much greater line efficiency is achieved (because no
start bits need be sent), but its more complex arrangements do increase the cost as
compared with asynchronous transmission
equipment. Byte synchronization is
established at the very beginning of the transmission or after a disturbance or line
break using a special synchronization (SYN) pattern, and only minor adjustments are
needed thereafter. Usually an entire user’s data field is sent between the synchronization
(SYN) patterns, as Figure 9.15 shows.
The SYN byte shown in Figure 9.15 is a particular bit pattern, used to distinguish it
from other user data.
9.10 HANDSHAKING
We cannot leave the subject of modems without at least a brief word about the Hayes
command set (nowadays also called the A T command set), for many readers will at
sometime find themselves faced with the question of whetheramodem is ‘Hayes
compatible’. By this, we ask whether it uses the Hayes command set, the procedure by
- FOR PROTOCOLS OF DATA 193
which the link is set up and the data transfer is controlled; if you like, the handshake
and etiquetteof conversation. Itallows, for example, aPC to instructa modem to diala
given telephonenumber and confirm when connection
the is ready. The Hayes
command set has become the de facto standard for personal computer communication
over telephone lines. An alternative scheme is today offered by ITU-T recommendation
V.25 bis.
9.11 PROTOCOLS FORTRANSFER OF DATA
Unfortunately the X.21, V.24/V.28 (RS-232) and RS449 standards and their inter-DCE
equivalents (i.e. the line transmission standards used by modems or digital links such as
V.22, V.32 or V.34) are still not enough in themselves to ensure the controlled flow of
data across a network. A number of additional layered mechanisms are necessary, to
indicate the coding of the data, to control themessage flow between the end devices and
to provide for error correction of incomprehensible information. Unlike human beings
engaged in conversation, machines can make sense at all of corrupted messages, and
no
they need hard and fast rules of procedure to be able to cope with any eventuality.
When they are given a clear procedure they are able, unlike most human beings, to
carry on several ‘conversations’ at once.These rules of procedureare laid out in
protocols.
Now, sit down and make yourself comfortable. . .
Once upon a time protocols were relatively unsophisticated like the simple computer-
to-terminalnetworks which they supportedand they were containedwithinother
computer application programmes.Thusthecomputer, besides itsmainprocessing
function, would be controlling the line transmission between itself and its associated
terminals and other peripheral equipment. However, as organizations grew in size and
data networks became more sophisticated and far-flung, the supporting communica-
tionssoftware andhardware developed to such an extent asto be unwieldy and
unmaintainable. Many computer items(particularly different manufacturers’equip-
ments) were incompatible. Against this background the concept of layered protocols
developed with theobjective of separating out the overall telecommunications functions
into a layered set of sub-functions, each layer performing a distinct and self-contained
task but being dependent on sub-layers. Thus complex tasks would comprise several
layers, while simple ones would need only a few. Each layer’s simple function would
comprise simple hardware and software realization and be independent of other layer
functions. In this way we could change either the functions or the realization of one
functional layer with only minimal impact on the software and hardware implementa-
tions of otherlayers. For example, a change in theroutingofa message (i.e.the
topology of a network) could be carried out without affecting the functions used for
correcting anycorrupteddata(orerrors)introducedonthe line between theend
terminals.
Most data transfer protocols in common use today use a stack of layered protocols.
By studying such a protocol stuck we have a good idea of the whole range of functions
that are needed for successful data transfer. We need to consider the functions of each
protocol layer as laid out in the international standards organization’s (ISO’s) open
- 194 PRINCIPLES
DATA NETWORK AND PROTOCOLS
systems interconnection (OSZ) model. The OSI model is not a set of protocols in itself,
but itdoescarefully define thedivision of functionallayersto which all modern
protocols are expected to conform. The protocols of each of the layers are defined in
individual I S 0 standards and ITU-T recommendations.
9.12 THEOPENSYSTEMS INTERCONNECTIONMODEL
The open systems interconnetion model, first standardized by I S 0 in 1983, classifies
data transfer protocols in aseries of layers. It sets worldwide standards of design for all
data telecommunicationprotocols,ensuringinterworkingcapabilityofequipments
provided by different manufacturers.
To understand the need for the model, let us start with an analogy, drawn from a
simple exchange of ideas in the form of a dialogue between two people. The speaker has
to convert his ideas into words; a translation may then be necessary into a foreign
language which can be understood by the listener; the words are then converted into
sound by nerve signals and appropriate muscular responses in the mouth and throat.
The listener meanwhile is busy converting the sound back into the original idea. While
this is going on, the speaker needs to make sure in one way or another that thelistener
has received the information, and has understood it. If there is a breakdown in any of
these activities,there can be nocertaintythatthe originalideahas been correctly
conveyed between the two parties.
Note that each function in our example is independent of every other function. It is
not necessary to repeat the language translationif the receiver did not hearthe message,
arequest (prompt)to replaya tape of thecorrectlytranslated message would be
sufficient. The specialist translator could be getting on with the next job so long as the
less-skilled tape operator was on hand. We thus have a layered series of functions. The
idea starts at the top the talker’s stack of functions, andis converted by each function
of
in the stack, until at the bottom it turns up in a soundwave form.A reverse conversion
stuck, used by the listener, re-converts the soundwaves back into the idea. Figure 9.16
shows our example.
Each functionin the protocol stack of the speaker hasan exactly corresponding, or so-
called peer function in the protocol stack the listener. The functions at the same
of layer
in the two stacks correspond to such extent that if we could conduct a direct
an peer-to-
peer interaction then we would actually be unaware of how the functions of the lower
layers protocols had been undertaken. Let us, for example, replace layers 1 and 2 by
using a telex machine instead. The speaker still needs to think up the idea, correct the
grammar and see to the language translation, but now insteadof being aimed at mouth
muscles and soundwaves, finger muscles and telex equipment do the rest, provided that
the listeneralsohasa telex machine. We cannot, however, simply replace only the
speaker’s layer 1 function (the mouth),if we do notcarry out simultaneouspeerprotocol
changes on the listener’s side because an ear cannot pick up a telex message. The prin-
ciple of layered protocols is that so long as the layers interact in apeer-to-peer manner,
and so long as the interface between the function of one layer and its immediate higher
and lower layers is unaffected, then itis unimportant how the function of that individual
layer is carried out. This is the principle of the open systemsinterconnection (OSZ)
- THE OPEN SYSTEMS INTERCONNECTION
MODEL 195
loyer
5 Ideo
Talker
Conversion t o
Ideo
Listener
1
8 L
longuoge,
grommor,
syntax
i s ready ond
Language
to ideo
conversion
Confirm receipt
(e.g. smile)
message (repeat
i f necessary)
1 Message to
mouth muscles
I Message t o
brain
1 Mouth
b Sound
L Eor I
Figure 9.16 A layered protocol model for simple conversation
model. The OS1 model sub-divides the function of data communication into a number
of layered and peer-to-peer sub-functions, as shown in Figure 9.17.
In all, seven layers are defined. Respectively, from layer seven to layer one these are
called: the application layer, the presentation layer, the session layer, the transport layer,
the network layer, the data link layer and the physical l q e r .
Each layer of the OS1 model relies on the service of the layer beneath it. Thus the
transport layer (layer 4) relies on the network service which is provided by the stack of
layers 1-3 beneath it. Similarly the transport layer provides a transport service to the
session layer, and so on.
The functionsof the individual layers of the OS1 model are defined more fully in I S 0
standards (IS0 7498), and in ITU-T’s X.200 series of recommendations; in a nutshell
they are as follows.
9.12.1 Application Layer(Layer 7)
This is the layer that provides communication services to suit all types of data transfer
between cooperating computers. It comprises a number of service elements (SEs), each
suitedforaparticularpurpose orapplication.They may be combinedinvarious
permutations to meet the needs of more complex applications.
A wide range of application layer protocols will be defined over time, to accommodate
all sorts of different computer equipment types, activities, controls other
and
applications.These will be defined in a modular fashion, the simplest common
functions being termed application service
elements (ASEs), which are sometimes
grouped in specific functional combinations to form application entities (AEs). These
are the functions coded as protocols.
- 196 PRINCIPLES
NETWORK DATA AND PROTOCOLS
OS1 layer
number
Peer - t o - peer
Application Application
protocol
Presentation c-----+ Presentation
S Session + - - - - - 4 Session
Transport c------@Transport
3 Network Network
Data link
1 Physical
l
c - - - - ) A c t u a lo m m u n i c a t i o n
- - - - - c
e- - -D I m a g i n a r yc o m m u n i c a t i o n( r e l y i n g on lower l a y e r s 1
Model. (Courtesy of CCZTT - derived
Figure 9.17 The Open Systems Interconnection (0%)
from Figures 12 and 13/X200)
Although the application
layer protocol will differ according to the particular
situation, AE or ASE, the I S 0 standards at leastprovideastandardized notation
(shorthand language)inwhichthe applicationlayerprotocols maybedefined and
written to enable fairly rapid interpretation by experienced technicians and equipment
designers. This notation is the abstract syntax notation I (ASN.1).It is laid out in ITU-T
recommendations X.208 (old version) and X.680 version). Examples of application
(new
layer protocols covered by this book are the transaction capabilities of signalling system 7
(see Chapter 12) and the message handling system (Chapter 23).
9.12.2 PresentationLayer(Layer 6)
As we found in Chapter 4, data may be coded in various forms such as binary, ASCII,
ITU-T IA5, EBCDIC, faxencoding, multimedia signal format, etc. The task of the
presentation layer is to negotiate a mutually agreeable technique for data encoding and
nguon tai.lieu . vn