1 Basic Knowledge

This chapter introduces briefly the basic knowledge of digital communication and SDH system for the sake of your understanding later chapters and sections free of trouble. If you have a good knowledge of SDH communication theories, please skip this chapter to later chapters and sections.

1.1 Brief introduction to communication network

1.1.1 Definition and classification of communication network
The communication system of multiple interconnected user communication systems is called a communication network. Communication networks can be classified in many ways, such as by purposes, by features of transmission signals, etc. y kinds of services implemented, communication networks can be categori!ed as telephone communication networks, data communication networks and broadcast television networks. y the network service scope, communication networks can be categori!ed as local networks, toll networks and international networks. y basic structural forms, communication networks fall into line network, ring network, star network, mesh network and hybrid network.

1.1.2 Basic elements of a communication network

The basic elements of a communication network are terminal e"uipment, transmission links, switching e"uipment and access part. The terminal e"uipment refers to the originating point and the terminating point on a communication network and includes some transform and inverse transform devices besides the devices corresponding to the signal source and signal sink. The terminal e"uipment varies for different communication services. #or e$ample% the terminal e"uipment of the telephone services is a telephone set and that of the data services is a data terminal. The transmission link is the connection medium between network nodes and the transmission channel for information and signals. esides corresponding to the channel part, the transmission link includes some transform and inverse transform devices, such as the open wire transmission, carrier transmission system, digital microwave transmission system, optical fiber transmission system, and satellite transmission system, etc. &ll of these can be used to implement communication network transmission link. The switching e"uipment is the core of a modern communication network. It is basically responsible for the concentration, transit and connection and distribution of the links connected to a switching node. Currently, the main mode is the circuit switched mode to connect a voice circuit, together with message switching mode and packet switching mode. The access part is the transmission entity between the service node interface and the user network interface, which connect the user to a service node via standard interfaces.

1.1.3 Development direction of communication networks

'ith the swift development of information process, the highly developed information society re"uires a communication network to provide various information services. To meet this demand, the modern communication network is now employing various computer(based intelligent terminal technologies and database technologies at an accelerated speed to develop toward a digital, integrated, intelligent and individual way.

1.2 Introduction to PCM3 !32 s"stem

1.2.1 Multiple#ing tec$nologies
In a communication network, the transmission medium costs far more than the e"uipment part does. Therefore, how to improve the line utili!ation ratio has been an important sub)ect in the communications field, which is a ma)or means to improve the line utili!ation ratio. &t present, the common multiple$ing technologies include #re"uency Division *ultiple$ing +#D*,, Time Division *ultiple$ing +TD*, and Code Division *ultiple$ing +CD*,. -ach channel of terminal signals in the transmission lines of the #D* system uses a different fre"uency band, that in the transmission lines of the TD* system uses a different time segment and that in the transmission lines of the CD* system uses a different coding mode. In this way, line multiple$ing is achieved.

1.2.2 Digitali%ation of voice signals

To implement TD* for analog voice signals that are continuous in time, the analog signals should be converted to signals that are discrete in time, namely, what we often call the digitali!ation of analog signals. .ulse Code *odulation +.C*, is a common technology of digitali!ation of analog signals. The simple block diagram of its communication system is shown in #ig. /(/. This is made up of three parts% +/, transmitting end, including low(pass filtering, sampling, "uanti!ation and encoding +namely, analog0digital conversion,1 +2, channel part, including transmission lines and regenerators1 +3, receiving end, including signal regeneration and digital0analog conversion which further includes decoding and low(pass filtering.

&nalog signal

4ow(pass filter

Sampling .C* Signal

5uanti!ation

-ncoding

6egenerator 4ow(pass filter

.C* signal

Transmit end +&nalog0 digital conversion)

6egeneration

Decoding

Channel &nalog 6eceive end signal ( Digtal0analong conversion )

#ig. /(/ The block diagram of the .C* system

In #ig. /(/, the basic units of .C* implements signals processing and performs the functions as follows% /. Sampling% The so(called sampling means sampling an analog signal value +sample value, at a

regular interval +T,. & series of sampled values that are discrete in time are called sample value se"uences. ased on the sampling theorem, the sample value se"uence can be recovered to the original analog signals without any distortion as long as the interval +T, between sample pulses is less than or e"ual to /02f m, i. e., the sampling fre"uency f s is more than or e"ual to 2f m +fm is the ma$imum fre"uency of the analog signal,. 2. 5uanti!ation% It means converting the sample signals in consecutive amplitude into sample value se"uences in discrete amplitudes by means of rounding(up. Thus, signal amplitudes can be represented with binary digital signals with finite bits. 3. -ncoding and decoding% -ncoding means converting the signal amplitude values sampled and "uanti!ed into a group of binary code elements. Decoding means restoring a group of binary code elements into the "uanti!ation values of corresponding signal amplitudes.

1.2.3 PCM 3 !32 &"stem

&fter the implementation of .C* digitali!ation of analog signals, TD* of multi(channel terminal signals can be further achieved. Split channels at the sample interval T into time segments, and the obtained time segment is called a frame. area frame is further split into 7 smaller time segments, each of which +T07, is called a time slot. 'ithin a frame, a time slot is allocated to the encoding information of each channel of terminal signals and multi(channel terminal signals are transmitted alternately, thus reali!ing .C* multiple$ing of channels. &lso, .C* multiple$ing is often adopted in a S.C digital switching system to increase the transmission rate and switching capacity. &ccording to IT8(T recommendations, the sample fre"uency of voice signals +399H!:3;99H!, is <kH!, the "uanti!ation level of sample values is 2=> and the digit of encoding bits of sample values is <, i.e, the transmission rate of single channel voice .C* signals is < ? <k @ >;kbit0s. Currently, there are two multiple$ing modes for .C* primary group% 39032(channel frame structure and 2;(channel frame structure. 39032(channel frame structure is adopted in China. That is, each frame occupies /2=As and is divided into 32 time slots, but only 39(channel voice encoding information is transmitted. The primary group rate when multiple$ed once is 32 ? >;kbit0s@29;<kbit0s@2.9;<*bit0s, namely, what we often call -/. -/ can be used to form a high order group or transmitted separately in such transmission channels as local call cables, toll cables, digital microwave, optical fibers, etc. The specific parameters and frame structures of .C*39032 primary group are as follows% /. asic characteristics 7umber of time slots0frame% 32 7umber of voice channels0frame% 39 Sample fre"uency% <kH! 7umber of encoding bits% n@< 7umber of "uanti!ation levels% *@2n@2=> 6ate of multiple$ing code stream% <999 32 <@29;<kbit0s #rame length% /2=As Single channel digital rate% >; kbit0s 2. #rame and multi(frame structures #ig. /(2 is the schematic diagram of the frame and multi(frame structures in .C* system, in which described in detail are the frame and multi(frame structures of and time slot allocation of .C* system.

/> #rame 2ms

*ultiframe

#9

#/

#2

#3

#;

#=

#>

#B

#<

#C

#/9 #// #/2 #/3 #/; #/=

/ #rame, 32 Timeslot, 2=> bit, /2= s TS

#rame

>

<

TS /9 // /2 /3 /; /= /> /B /< /C 29 2/ 22 23 2; 2= 2> 2B 2< 2C 39 3/ CH/> CH39

Signaling timeslot

#rame alignment timeslot

CH/ CH/=

-ven #rame Ddd #rame

9 9 / / 9 / / #rame alignment signal / & / / / / / /

#9

9 9 9 / &2 / / .olarity code b c d CH/> b c d CH/B b c d CH39

/ 2 3 ; = > B < Segm ent cde Segm ent cde

*ultiframe alignment code

#/

b c d a CH/

#2

b c d a CH2

#/=

b c d a CH/=

7ote% +/, &nd &2 are respectively remote end alarm codes during out(of(frame and multiframe out( of(syn. 7ormal% 9. &larm% /. +2, 'hen b, c and d codes are not used, it is always b@/, c@9,d@/. +3, is reserved for international communication and stipulated as / temporarily.

#ig. /(2 The schematic diagram of the frame and multiframe structures in .C* system

3. Time slot allocation In 39032 mode, the sample cycle is /0<999@/2=As, i. e, /2=As is a frame. There is 32(channel TD* in a frame and each channel occupies as many as /2=032@3.CAs time slots. There are 32 time slots in a frame, which are numbered in se"uence as TS9:TS3/. The time slots are allocated and used as follows% TS/:TS/= and TS/B:TS3/ are 39 voice channel time slots. TS9 is frame synchroni!ation code and monitor code time slot. TS/> is signaling +ring, busy line and other identification signals, time slot. ;. &rrangement of voice channel bits 'ithin each voice channel time slot, the sample values should be encoded as binary codes and each code element occupies 3. CAs0<@;<<ns called / bit, numbered as /:<. =. &rrangement of bits in TS9 time slot To keep strict synchroni!ation between the transmitting and receiving ends, each frame should transmit a group of frame synchroni!ation codes or monitor codes with special identifiers. -ven frame TS9 is a frame synchroni!ation code group% E99//9//, the first code bit F is used for international communications and the codeG/Hwill be sent when this code group is not used. The bits of the odd frame TS9 are allocated as% The third code bit is used for alarm of out(of(sync. and is represented with &/. The code G9Hshould be sent during synchroni!ation and G/H during out(of( synchroni!ation. To prevent 2:< bits of the odd frame TS9 from becoming dummy synchroni!ation codes, the second bit is specified as the monitor code, which is fi$ed as G/H. The bits from the ;th to the <th are used internally and specified as G/H for the moment. >. &rrangement of bits in TS/> time slot If the code bits of TS/> are allocated in the order of time to various voice channels for transmitting signaling, /> frames are needed to form a multi(frame, represented respectively with

#9:#/=. The fre"uency of this multi(frame is =99H! and the cycle is 2ms. TS/>s of various sub( frames in this multi(frame are allocated as follows% #9 frame% /:; code bits transmits the multi(frame synchroni!ation signal 9999. The > th code bit transmits remote alarm signal &2 for multi(frame being out of sync., which is G9H during synchroni!ation and G/H during out(of(sync. =, B and < code bits transmit the codeG/H. #/:#/= frames% The first ; bits of TS/> in each frame transmit CH/:CH/= signaling signals and the last ; bits transmit CH/>:CH39 signaling signals.

1.3 'verview of optical fi(er communication

Dptical fiber is short for optical waveguide fiber. Dptical fiber communication is a communication mode in which optical signals are used as carriers and optical waveguide fibers as transmission media. ecause optical fiber communications is characteri!ed by a series of advantages such as broad transmission fre"uency band, high communication capacity, low loss and free from electromagnetic interference, the optical fiber communication technology has been developing at full speed in recent years.

1.3.1 Division of optical wave (ands

Dptical waves are electromagnetic waves the most familiar to human beings, and the wavelength is at the micron level and the fre"uency at the /9 /;H!:/9/=H! "uantity level. Currently, the wavelength range for optical fiber communications is near infrared region, i.e., 9.<Am:/.<Am. Dptical waves can be classified as the waveband with short wavelength of 9. <=Am and waveband with long wavelength of /.3/Am and /.==Am. These three wavelengths are currently used in practical optical fiber communication. -ach wavelength range for optical transmission is also called an optical transmission window.

1.3.2 &tructure of optical fi(er

Dptical fibers currently used for communication are concentric double(layer cylinders in small cross(sections made from "uart! glass +SiD 2,, which are called bare optical fibers when not coated or without plastic sleeve. &s shown in #ig. /(3, a bare optical fiber is made up of fiber core and cladding. The central part in a high refractive rate as n / and with 2a diameter is called fiber core, while. The central part in a low refractive rate as n 2 is called the cladding, with the diameter of 2b. & and b are respectively different values since they transmit optical signals with different wavelengths and in different modes on optical fiber. ecause "uart! glass is brittle and easy to break, the surface of bare optical fiber should be coated twice to form optical fiber core so as to protect the surface of optical fiber and increase tensile strength in use. &s shown in #ig. /(;, an optical fiber core cable is made up of fiber core, cladding, coating and plastic sleeve. The outside of a cladding is coated with a thin layer with silicone resin or polyurethane. The outer plastic sleeve of a coating +or called second coating, is mostly of such plastic as nylon or polythene.

2a 2b 2a 2b

#ig. /(3 The schematic diagram of lateral section structure of a bare optical fiber
#iber core Cladding

#irst coating

Second coating +.lastici!er,

2a 2b 2a 2b

#ig. /(; The schematic diagram of lateral section structure of an optical fiber core cable

1.3.3 Classification of optical fi(ers

Dptical fibers can be classified in terms of optical fiber materials, manufacturing processes, transmission modulus, distribution of refractive rate on cross section of optical fiber and operating wavelength. 5uart! optical fiber currently adopted in communication is often categori!ed in the following two aspects% /. Classification by different refractive rates +/, 8niform optical fiber. The refractive rate n / of the optical fiber core and the refractive rate n 2 of the cladding are both constants and n /In2, and the refractive rate at the edge between a fiber core and a cladding changes by step. This kind of optical fiber is called a uniform optical fiber. +2, 7on(uniform optical fiber. The refractive rate n / of an optical fiber core decreases regularly as the diameter increases and e"uals the refractive rate n 2 at the edge between the fiber core and the cladding. This kind of optical fiber is called a non(uniform optical fiber. 2. Classification by the number of transmission modes 'hat we term a mode is a distribution form of electromagnetic field in essence. Different modes have different electromagnetic distribution forms. In terms of the number of transmission modes in optical fiber, there are single mode optical fibers and multimode optical fibers.

+/, & single mode +S*, optical fiber% its fiber core has a diameter, usually ;Am:/9Am. It transmits only one mode theoretically. The single mode optical fiber transmits main mode only, thus avoiding mode dispersion and making this kind of optical fiber with a broad transmission fre"uency band. So it features high transmission capacity, and therefore, applicable to optical fiber communication of high capacity and long distance. +2, *ultimode +**, optical fiber% when an optical fiber has multiple modes transmitted at a certain operating wavelength in it, this kind of optical fiber is called a multimode optical fiber. The fiber core of a multimode optical fiber is generally =9Am:B=Am in diameter and the cladding is /99Am:299Am in diameter. This kind of optical fiber has poor a transmission performance, a narrow bandwidth and a low transmission capacity. ecause a single(mode optical fiber features great bandwidth, easy upgrade and e$pansion and low cost, it is considered internationally that only single mode optical fiber should be used as the transmission medium in the synchronous optical cable digital transmission system. In the 3 optical transmission windows, <=9nm window is used for multimode transmission only and /3/9nm and /==9nm windows for single mode transmission. How far the optical signals are transmitted in an optical fiber will be influenced by both dispersion and loss. Dispersion will e$tend digital pulses transmitted in the optical fiber, thus causing inter( symbol interference and reducing signal "uality. 'hen inter(symbol interference deteriorates transmission performance to some e$tent, the transmission system will function no more. 4oss makes the intensity of the optical signal transmitted in an optical fiber decrease gradually as the transmission distance increases. 'hen the optical power is reduced to some e$tent, the transmission system will function no more. To e$tend the transmission distance of the system, people focus their ma)or attention on reduction in dispersion and loss. The /3/9nm optical transmission window is called a !ero dispersion window because optical signals have the lowest transmission dispersion in this window. The /==9nm window is called the minimum loss window because the optical signals have the minimum transmission attenuation in this window. IT8(T specifies three common optical fiber specifications% J. >=2, J. >=3 and J. >=;. J.>=2 optical fiber has the best dispersion performance in the /3/9nm wavelength window and is also called optical fibers without any shift dispersion +i. e. the !ero dispersion window is located at /3/9nm wavelength,. J.>=2 optical fiber is applicable to two wavelength areas /3/9nm and /==9nm. J.>=3 optical fiber means the single mode optical fiber with the best dispersion performance in the /==9nm wavelength window and is also called the single mode with shift dispersion. It shifts the !ero dispersion point from /3/9nm to /==9nm by changing the distribution of refractive rates within an optical fiber to achieve reduction in dispersion and loss in the /==9nm wavelength window. This kind of optical fiber is mainly used in the /==9nm operating wavelength area. J.>=; optical fiber is called the minimum loss optical fiber in the /==9nm wavelength window and its !ero dispersion point is still located at the /3/9nm wavelength. This kind of optical fiber mainly works in the /==9nm window and is chiefly applicable to seafloor optical fiber communications which needs a very long regenerator section transmission distance.

1.) &"nc$ronous Digital *ierarc$" +&D*,

1.).1 Brief Introduction to &D*
The full name of SDH is Synchronous Digital Hierarchy. SDH specifies such features as the frame structure, multiple$ing mode, level of transmission rate, interface code pattern of digital signals and provides an internationally supported framework. &nd based on this framework, it can be developed into a fle$ible, reliable and manageable world telecommunication network. This transmission network is easy to e$tend and applies to the development of new Telecom services. esides, it makes it possible for the e"uipment from different manufacturers to interconnect with each other, which is )ust what the network constructors have long been e$pecting.

1.).2 Disadvantages of PD* and emergence of &D*

efore SDH , it is the .lesio(chronous Digital Hierarchy +.DH, that is used in the transmission system. .DH is a multiple$ing technology that multiple$es signals of a low rate level into the high(speed ones by adopting bit filling and code bit interleaving. It can separately transmit domestic toll and local network services. #or e$pansion, it is only necessary to add new .DH e"uipment.. However, with the development of Telecom network and higher user demands, the intrinsic disadvantages of .DH are gradually revealed% /. It has only the regional digital signals and frame structures but no international standards. There are currently three international signal rate levels, namely, -uropean hierarchy, 7orth &merica hierarchy and Kapan hierarchy. 7orth &merica and Kapan adopt the /.=* system while -urope 2* system, which is also adopted in China. -uropean rate standards are 2*bit0s +-/,, <*bit0s +-2,, 3;*bit0s +-3, and /;9*bit0s +-T;,. The rate standards of 7orth &merica are /.=*bit0s +T/,, >.3*bit0s +T2, and ;=*bit0s +T3,, and those of Kapanese system are /.=*bit0s, >.3*bit0s and 32*bit0s. The three current signal rate levels are incompatible with one another, thus making international network more difficult. 2. 4ack of worldwide standard optical interface specifications makes the special optical interfaces developed by various manufacturers different from each other and incompatible with each other. Thus, internetworking fle$ibility is restricted, along with increased comple$ity and operation costs. 3. .DH is a multiple$ing structure based on point(to(point transmission. That is, it only supports point(to(point transmission to form a line network segment by segment. eside, it can only implement segment protection but not the ring protection of various routes operating on a unified basis. Therefore it features infle$ible network topology, low utili!ation of digital e"uipment and has no optimal routing. ;. The operation, management and maintenance of traditional .DH rely mainly on manual digital signal cross(connection and tests with services stopped. Therefore, no bit is arranged in the multiple$ing signal frame structure for the purpose of the running, management and maintenance of a network. &nd such lack of overhead bit makes it difficult to establish centrali!ed transmission 7* and hard to satisfy the re"uirements of users for dynamic network and access of new services. =. In the multiple$ structure of .DH, only the signals at the low rate levels like 7orth &mericaLs /.=*bit0s, Kapanese /.=*bit0s and >.3*bit0s and -uropean 2*bit0s are multiple$ed in synchronous mode, those of other levels are multiple$ed in asynchronous mode. That is, some

additional bits are filled to synchroni!e various tributary signals and multiple$ them into high( speed signals. ut in this mode, it is difficult to recogni!e and e$tract low(speed tributary signals from high(speed signals. To add0drop a voice channel, the only way is to de(multiple$ the whole high(speed line signals step by step to low(speed line signals to be dropped. &fter a voice channel is added0dropped, the low(speed line signals are to be multiple$ed step by step into high(speed line signals for transmission. #or e$ample, to branch out a 2*bit0s low(speed tributary signal from the /;9*bit0s code stream by adopting .DH, first optical signals are converted into electric ones, which are to be de(multiple$ed thrice, namely, /;9*bit0s 3;*bit0s +/;9* de(multiple$ed into 3;*,, 3;*bit0s<*bit0s and <*bit0s2*bit0s to drop a voice channel. Then, the electric signals are to be multiple$ed thrice, namely, 2*bit0s <*bit0s +2* multiple$ed into <*,, <*bit0s3;*bit0s and 3;*bit0s/;9*bit0s for transmission +refer to #ig./(=,. It can be seen that the .DH system not only has comple$ and infle$ible multiple$ structure, but also a large amount of hardware, high add0drop service costs and comple$ implementation of digital cross( connection function. To satisfy the development re"uirements of a modern Telecom network and usersL service demands, the best solution is to reform the technical system fundamentally because solving the above problems within the original system and technical framework would be half the result with twice the effort. SDH came into being as a new system, which combines high(speed and high( capacity optical transmission technology and intelligent network technology.

1.).3 -dvantages of &D*

ecause SDH came into being )ust to overcome the disadvantages of .DH, first the ob)ective, then the specifications are specified and the e"uipment is developed. This process is right contrary to that of .DH. Dbviously, this makes it possible for people to define the system and e"uipment conformable to future Telecom network re"uirements in the optimal mode and to the greatest e$tent. The following main features of SDH represent these re"uirements. /. 8nify the regional standards of 7orth &merica, of Kapan and of -urope at the ST*(/ level or higher. Therefore, digital signals in communication crossing countries do not have to be converted into another standard, and the worldwide standards for the digital transmission system are reali!ed for the first time in a real sense. 2. Since there is a unified standard optical interface, hori!ontal compatibility can be reali!ed on basic optical cable sections. That is, the inter(network the e"uipment of different manufacturers is enabled on optical channels to satisfy the re"uirements of manufacturers for environment. 3. SDH adopts synchronous multiple$ing mode and fle$ible multiple$ mapping structure and the code streams at various levels are arranged regularly within the frame structure payload. &nd because the payload is synchronous with the network, it is only necessary to use the software to add0drop high(speed signals into low(speed tributary signals at one time, which is the so(called one(step multiple$ing characteristic. .lease compare the process of add0drop signals in SDH and that in .DH by referring to #ig. /(=% To add a 2*bit0s low(speed tributary signal from the /==*bit0s code stream, you have to first use the add0drop multiple$er +&D*, of SDH and then software to directly add the 2*bit0s tributary signal at one time. In this way, the process of hierarchically de(multiple$ing and re(multiple$ing all high(speed signals is avoided and the whole set of back(to(back multiple$er is not necessary any more. Therefore, the services of SDH are easy to add0drop, the network architecture and e"uipment are simplified to a great e$tent and digital

cross(connection is very easy to implement.

/;903;*bit0s Dptical0 electric Dptical0 electric .DH
Demultiple$

3;0/;9*bit0s

3;0<*bit0s <03;*bit0s
Demultiple$ *ultiple$

*ultiple$

-lectric0 optical

Demultiple$

<02*bit0s

Demultiple$

*ultiple$

20<*bit0s

2*bit0s +-lectric signal,

SDH
/==*bit0s Dptical interface Dptical interface 2*bit0s +-lectric signal, &D* /==*bit0s

#ig. /(= Comparison between the adding0dropping signals of SDH and those of .DH

;. & large amount of software is adopted in SDH for network configuration and control, which makes configuration more fle$ible and dispatching more convenient. =. Diversified overhead bits are arranged in SDH frame structure. These overhead bits cover about =M of all signals and can be processed with software, which greatly enhances the network capabilities in terms of its running, management and maintenance. >. The SDH network is completely compatible with the current network. That is, SDH is compatible with all the rates of current .DH, which enables SDH to support the .DH network already constructed and facilitates the smooth transition from .DH to SDH. esides, the SDH network accommodates new service signals such as &T* cell, i. e., SDH has complete backward compatibility and forward compatibility.

1.).) &D* rate

The rate level of SDH signals is represented as ST*(7 +7 is an integer,. Currently, SDH supports certain 7 values only, i.e., 7 can only be /, ;, /> and >;. The most basic and most important module signal is ST*(/, with the rate being /==. =29*bit0s. The ST*(7 signal at a higher level is obtained by means of byte interleaving of the basic module signal ST*(/. The rates at the ST*(; level, at the ST*(/> and at the ST*(>; level are respectively >22. 9<9*bit0s, 2;<<, 329*bit0s and CC=3. 2<9*bit0s.

1.).. &D* frame structure

The frame structure is shown in #ig. >. 2(>.

C $ 7 Column + yte,

2>/ $ 7 Column + yte,

/ 6egenerator section overhead +6SDH, &dministrative unit pointer +&8(.T6, *ultiple$ section overhead +*SDH, C /2= s 2B9 $ 7 Column ST*(/ .ayload +.ayload,

Transmission direction
C 6ow

3 ; =

#ig. /(> SDH frame structure

SDH transmission is in the unit of byte and its frame structure is a rectangular massive one based on the byte structure, including 2B9 7 columns and C(row bytes, each of which has < bits. The rectangular frames of SDH are transmitted row by row on optical fibers after parallel0serial conversion at the optical transmitting end and are recovered into rectangular massive ones for processing after serial0parallel conversion at the optical receiving end. The bytes in an SDH frame are transmitted row by row from left to right, beginning with the first byte at the left top of the figure. This transmission row by row continues until C 2B9 7 bytes are all transmitted. Then it is time for the ne$t frame to be transmitted. Thus, one frame after another is transmitted. <,999 frames with the constant frame length being /2=As can be transmitted. The frame fre"uency of SDH is <,999 frames0second. That is, a specific byte in a signal frame is transmitted <,999 times per second and the bit rate of this byte is <,999 <bit@>;kbit0s, namely, the transmission rate of /( channel digital telephone. Take for e$ample the ST*(/ level, whose rate is 2B9 +2B9 columns0frame, C +altogether C rows, >;kbit0s +>;kbits for each byte, @/===29kbit0s @/==. =29*bit0s. It can be seen in #ig./(> that the frame structure of ST*(7 is made up of three parts% section overhead, including 6egenerator Section Dverhead +6SDH, and *ultiple$ Section Dverhead +*SDH,, Information .ayload +.ayload, and &dministrative 8nit .ointer +&8(.T6,. /. Section Dverhead +SDH, area SDH means the additional bytes in the ST*(7 frame structure needed for normal and fle$ible transmission of information payload and these bytes are mainly used for the running, management and maintenance of the network. In the /: C 7 columns of the SDH frame, /:3 rows and =:C rows are allocated to the SDH. SDH can be further categori!ed as 6SDH and *SDHN. /:3 rows are allocated to 6SDH and =:C rows to *SDH. 6SDH can be accessed either at the regenerator to at the terminal e"uipment. However, *SDH passes a regenerator transparently and is terminated at the terminal e"uipment. 2. .ayload +.ayload, area Information payload area is the place where information about various services is stored in the SDH frame structure. Hori!ontal columns /9 7:2B9 7, and vertical rows /:C belong to the

information payload area. In it, there are still some .ath Dverhead +.DH, bytes transmitted as part of the payload in a network and these bytes are mainly used for the monitor, management and control of the path performance. 3. &dministrative 8nit .ointer +&8(.T6, area &8 .T6 is a kind of indicator, mainly used to indicate the accurate position of the first byte of information payload in the ST*(7 frame, so that the information can be correctly decomposed at the receiving end. It is located at the fourth row of /:C 7 columns in the ST*(7 frame structure. The adoption of the pointer mode is an innovation of SDH. It can perform multiple$ synchroni!ation and ST*(7 signal frame locating in the "uasi(synchroni!ation environment.

1.)./ &D* multiple# mapping structure and multiple# mapping process

IT8(T specifies a whole set of multiple$ing structure +as shown in #ig. /(B, and via these lines three series of digital signals of .DH can be multiple$ed into ST*(7 signals in many ways. China specifies a very simple multiple$ mapping structure +as shown in #ig. /(<, to make each payload have only one multiple$ mapping path and this structure is a subset of standard multiple$ mapping structures specified in the specifications.

STM -N

A U G

A U -4

V C -4

C -4 TU G -3 TU -3 V C -3 C -3

1392 4!b" t # $

A U -3

V C -3

44%3 !b" t # $ 343 &!b" t # $ 312!b" t # $ 2'4&!b" t # $ 1(44!b" t # $

.ointer processing *ultiple$ 4ocating alignment *apping

TU G -2

TU -2 TU - 12 TU - 11

V C -2 V C - 12 V C - 11

C -2 C - 12 C - 11

#ig. /(B SDH multiple$ structure specified in IT8(T

Three steps mapping, locating alignment and multiple$ must be taken to load various signals to the SDH frame structure. *apping is e"uivalent to the binding of signals and synchroni!es different tributary signals and corresponding Oirtual Container, level n +OC(n,. 4ocating alignment means adding )ustification pointer and is used for correction of tributary signal fre"uency difference and implementation of phase alignment. *ultiple$ means byte interleaving multiple$ and is used for adaptation of multiple low order path layer signals to high order paths or multiple high order path layer signals to a multiple$ section layer.

STM -N

A U G

A U -4

V C -4
3 / 7

C -4

1392 4!b" t # $

TU G -3

TU -3 TU G -2

V C -3

C -3

343 &!b" t # $

.ointer .rocessing *ultiple$ 4ocating alignment *apping

TU - 12

V C - 12

C - 12

2'4&!b" t # $

#ig. /(< SDH multiple$ structure specified in China

#irstly, digital streams at various rate levels enter corresponding interface containers C, which are an information structure used to load various rate service signals and perform mainly the function of adaptation +such as rate )ustification,. Thus, the most commonly used plesiochronous digital system signals can be made to enter finite standard containers and perform such an adaptation function as rate )ustification. #or e$ample, for the 2* signals from various channels, some may be 2.9;</*bit0s while others may be 2.9;<2*bit0s because there are different clock accuracies. These will receive error tolerance )ustification in C so as to be adapted to standard signals at consistent rates. Currently, there are = kinds of standard containers% C(//, C(/2, C(2, C(3 and C(;. &s defined in China, C(/2 corresponds to the rate 2.9;<*bit0s, C(3 to the rate 3;.3><*bit0s and C(; to the rate /3C.2>;*bit0s. The digital streams out of standard containers plus .DH forms a OC and this process is mapping. OC is the most important information structure in SDH and mainly supports path layer connection. The OC enveloping rate synchroni!es with the network synchronous, so the enveloping of different OCs is synchronous, but inside the envelope, different(capacity plesiochronous tributary signals are allowed. -$cept at the OC assembling and dissembling points +i.e., the .DH0SDH network margins,, the transmission of OC in the SDH network remains always unchanged. So a OC can be taken as an independent entity to be inserted or taken out at any point on the channel, to make the synchronous multiple$ing and cross(connect, which is very convenient and fle$ible. OC can be classified as low order OC and high order OC. OC(/2 and OC( 3 here are low order OCs while OC(; is a high order OC +OC(3 in &8(3 is a high order OC. OC(3 should be included in low order OCs if adapted to OC(; via T8(3,. The digit flow from OC goes along the route as specified in the diagram and enters the administration unit +&8, or tributary unit +T8,. In an SDH frame, OC(n is a separate whole and cannot be split and transmitted. Therefore, the conversion of OC(n into T8(n and that of OC(n into &8(n are the process of rate adaptation, namely, that of locating alignment in a multiple$ structure. &8 is an information structure providing adapting functions for the high order channel layer. It consists of the high order OC and &8 .T6. Here, &8 .T6 is used to specify the position of the high order OC in the ST*(7 frame, so the high order OC in ST*(7 can float, but the position of &8 .T6 itself in ST*(7 shall be fi$ed. Dne or more &us occupying a fi$ed location in the ST*( 7 frame form an &dministrative 8nit Jroup +&8J,, which is made up of three &8(3s or a single &8(; in the byte interleave mode. Similarly, T8 is an information structure that provides adapting functions for the low order channel layer and high order channel layer. It consists of the low order

OC and T8 .T6. T8 .T6 is used to identify the location of a low order OC in the frame structure. Dne or more T8 with fi$ed positions in the high order OC payload form the tributary unit group +T8J,. In the end, 7 &8Js plus additional SDH forms the final ST*(7 frame structure. 7ow take for e$ample the multiple$ mapping process of the 2* tributary signals to detail the above. The signals with the nominal rate being 2.9;<*bit0s first enter C(/2 for adaptation processing and C(/2 plus .DH, when mapped, forms OC(/2. &fter locating alignment, the .T6 in T8(/2 identifies the phases of OC(/2 relative to T8(/2. 3 T8(/2s are multiple$ed into T8J(2 after even byte interleave and B T8J(2s into T8J(3 in the same way. 3 T8J(3s after byte interleave plus high order .DH forms OC(; payload and that after locating alignment plus .T6 forms &8(;. & single &8(; is directly put into &8J. In the end, 7 &8Js after byte interleave plus SDH forms ST*(7 signals.

1.. &D* over$ead function

1...1 'ver$ead t"pe
There are two different kinds of overhead arranged in the SDH frame structure, namely, SDH and .DH, used respectively for the maintenance of the section layer and path layer. That is, the overhead in the SDH system is used hierarchically. /. SDH SDH includes framing information, information used for maintenance and performance monitor and other operating functions. SDH can be further categori!ed as 6SDH and *SDH. 6SDH can be accessed either at the regenerator, or at the terminal e"uipment. *SDH passes a regenerator transparently and can only be terminated at the combination and termination points of &8J, namely, terminal e"uipment. In the SDH frame structure, /:3 rows are allocated to 6SDH while =:C rows to *SDH. 2. .DH .DH can be classified as two types% low order OC .DH and high order OC .DH. +/, 4ow order OC .DH% 'hen the low order OC .DH is attached to C(/0C(2, OC(/0OC(2 can be obtained. Its main functions are the monitoring of OC path functions, transmission of maintenance signals and alarm status indication. +2, High order OC .DH% 'hen OC(3 .DH is attached to C(3 or the combination of multiple T8J(2s, OC(3 can be obtained. 'hen OC(; .DH is attached to C(; or the combination of multiple T8J(3s, OC(; can be obtained. The main functions of high order OC .DH are the monitoring of OC path functions, transmission of maintenance signals, alarm status indication and multiple$ structure indication.

1...2 -rrangement of section over$ead

/. 4ocation of section overhead bytes Oarious SDH bytes are arranged in the ST*(/ frame as shown in #ig. /(C.

C A 1 B 1 D 1 A 1 A 1 A 2 E1 D 2

yte A 2 A 2 JO F1 D 3 6SDH

&dministrative 8nit .ointer

B 2 D 4 D D 1# $ 1 B 2 B 2 K1 D 5 D ! D 11 % 1 K2 D 6 D " D 12 E2 *SDH

Characteristic bytes related to transmission media +used temporarily, ytes reserved for domestic use 8nscrambled bytes used domestically &ll unlabelled bytes are to be specified in compliance with international strandards +medium(related applications, additional domestic uses and other uses,

#ig. /(C

yte arrangement in ST*(/

In SDH, the SDH frame in an ST*(7 frame are made up of 7 SDHs in the ST*(/ frame arranged in interleaved mode. However, only the first SDH of ST*(/ is reserved completely and other 7(/ SDHs of the ST*(/ frame have only the framing bytes &/, &2 and 2 reserved, with all the other bytes ignored. The location of one SDH byte in an ST*(7 can be represented with the coordinate vector S +a, b and c,, in which a stands for the number of rows, with the value range being /:3 or =:C. b stands for the number of multiple columns, with the value range being /:C. c stands for the number of interleave layers in a multiple$ column, with the value range being /:>;. The relation between the actual number of rows, that of columns of this byte in an ST*(7 and a, b, c is as follows% number of rows @ a number of columns @ 7 +b(/, P c 2. #unctions of SDH +/, #raming bytes% &/ and &2 The bytes &/ and &2 in SDH are used to identify the initial location of a frame. &/ and &2 have fi$ed binary values, i.e., &/ is ////9//9 and &2 99/9/999. If some = consecutive frames cannot receive correct &/ and &2 bytes, i.e., the = consecutive frames cannot distinguish among different frames, then the transmitting end enters the status of out of frame and generates an out(of(frame +DD#, alarm. If DD# continues for 3ms, loss of frame will occur and the e"uipment generates a 4oss of #rame +4D#, alarm. In this case, if an &larm Indication Signal +&IS, is inserted, the whole service will be interrupted. In the status of 4D#, if the receiving end is in the status of framing for consecutive /ms, the e"uipment will recover to normal. +2, 6egenerator section trace byte% K9 This byte is used to repeatedly send section access point identifiers so as to enable the section receiving end to confirm that it and the specified transmitting end are in continuous connection status. In the network of the same operator, this byte can be any character while the K9 bytes at the

transmit0receiving ends must be matched at the network boundary between two different operators. The K9 bytes enable operators to find and eliminate faults in advance and shorten the network restoration time. The K9 byte has another use. The K9 byte of each ST*(/ frame in the ST*(7 frame is defined as the identifier C/ of ST*, which is used to indicate the location of each ST*(/ in ST*(7. That is, the K9 byte indicates what ST*(/ +number of interleave layers, in ST*(7 this ST*(/ is and what column +number of multiple$ columns, in this ST*(/ frame this C/ is. +3, Data Communication Channel +DCC,% D/:D/2 DCC in SDH is used to form the transmission link of SDH *anagement 7etwork +S*7,. Though there are also control channels in a traditional plesiochronous system, they are all private and cannot be accessed e$ternally. However, DCC is universal and is embedded in SDH. esides, all network units have DCCs, which makes it easy to form a unified management network and saves many devices. The bytes D/:D3 are called 6S DCC, which is used for transmission of D&* +operation, maintenance and management, information at the rate of /C2kbit0s +3 >;kbit0s, between any two 6S terminals. However, the bytes D;:D/2 are called *S DCC, which is used for transmission of D&* information at the rate of =B>kbit0s +C >;kbit0s, between any two *S terminals. The whole B><kbit0s DCC provides a powerful communication structure for the management and control of an SDH network. #or e$ample, one important ob)ective of SDH network management Q control is to implement fast distributed control. 'ith DCC, the optimal route table calculated by the 7*S can be transmitted fast to a network unit at any time. +;, Drder wire byte% -/ and -2 These two bytes are used to provide the order wire communication voice channel. -/ belongs to 6SDH and is used for local order wire channel. -/ can be accessed at a regenerator. -2 belongs to *SDH and is used for direct order channel. -2 can be accessed at an *S terminal. The rate of an order wire is >;kbit0s. +=, 8ser channel% #/ This byte is reserved for usersL +often referring to a network provider, private use and can be used to provide temporary data0voice connection for special maintenance purposes. +>, it Interleave .arity <(bit Code + I.(< code,% / / byte +< bits, is used for the monitoring of regenerator section error code and is bit interleave parity code using even parity check. I.(< code calculates all the bits of the first ST*(7 frame after scramble and the results are located at the / byte before scramble. This kind of error code monitor is one feature of SDH and implements automatic error code monitor of a regenerator section in a very simple way. However, this method does not function when even error codes happen to occur in the same monitor code group. #ortunately, there is little probability of such a case, therefore the overall error code detection probability is very great. +B, it Interleave .arity 7 2;(bit code + I.(7 2;(bit code,% 2 2 byte is used for the monitoring of multiple$ section error codes and there are 3 2 bytes +altogether 2; bits, arranged for this purpose. 2 byte is bit interleave parity 7 2;(bit code using even parity check and its mode of generation is similar to I.(<. I.(7 2; calculates all the bits of the first ST*(7 frame +e$cept for /:3 rows in SDH, and the results are located at the 2 byte before scramble. There are in the ST*(7 frame 7 3 2 bytes, each of which corresponds to one parity code of the ST*(/ frame. In addition to the / bytes and the 2 bytes in the regenerator section and multiple$ section

arranged for error code monitor, SDH arranges one 3 byte at the OC(30OC(; high order channel layer .DH and the first and second bits at the OC(/0OC(2 low order channel layer .DH for error code monitor. It can be seen that SDH is very complete in terms of error performance monitor. -ach layer of network has performance monitor of four different hierarchies and can implement error code monitor of both as small as a regenerator section and as big as a OC(/0OC(2 channel. +<, &utomatic .rotection Switching +&.S, channel% R/ and R2 R/ and R2 bytes are used for the &.S signaling of multiple$ section protection and can gain a great protection response speed because they are embedded signaling channels used specially for protection purpose. R/ and R2 provide the network protection mode, whose operating principles are briefly described as follows% 'hen a fault occurs to some operating channel, a downlink end will detect the fault soon and use the uplink protection optical fiber to send the R/ byte containing the fault channel number. The uplink end receives the R/ bytes and connects the optical fiber bridge of local downlink operating channel with the downlink protection optical fiber. esides, the uplink end uses the downlink protection optical fiber to send some protection bytes R/ and R2, with the R/ bytes used for switching and the R2 bytes for acknowledge. The downlink end receives the R2 bytes and confirms the channel numbers. esides, the downlink end implements the bridge connection of the downlink operating channel with the downlink protection optical fiber at local end, and that of the uplink operating channel with the uplink protection optical fiber at local end as re"uired by the R/ bytes. To meet the re"uirements for bi(directional switching, the downlink end sends the R2 bytes via the uplink protection optical fiber. The uplink end receives the R2 bytes and implements the bridge connection of the uplink operating channel with the uplink protection optical fiber at local end. Thus, two pieces of operating channel optical fiber are switched to two pieces of protection optical fiber almost at the same time and the &.S is over. +C, Synchroni!ation status% S/ +b=:b<, In the ST*(7 frame structure, the =:< bits of the first S/ byte +C,/,/, in the first ST*(/ frame stand for synchroni!ation status information and these four bits may have /> different kinds of codes representing /> different synchroni!ation "uality grades. The smaller the value of S/ +b=:b<, is, the lower the corresponding clock "uality grade is. Thus, the e"uipment can evaluate the "uality of the received clock signals and decides whether to implement clock source switching, namely, switching to a clock source of higher "uality. +/9, *ultiple$ Section 6emote -rror Indication +*S(6-I, byte% */ */ byte is remote alarm information and is sent back by a receiving end to a transmitting end. The contents of */ byte is the number of error blocks detected by the receiving end via I.(7 2; + 2, code to enable the transmitting end to know about the receiving error code of the receiving end. +//, Special bytes related to transmission media These bytes are used for special applications related to transmission media, such as early alarm of protection switching in microwave SDH, automatic sending power control, fast non(traumatic switching control, propagation monitor, etc.

1...3 -rrangement of $ig$ order c$annel over$ead

/. 4ocation of high order channel overhead bytes The frame structure of OC(3 is a rectangular structure made up of C rows and <= columns, of

which the C bytes of the first column are used as OC(3 .DH. The frame structure of OC(; is a rectangular structure made up of C rows and 2>/ columns, of which the C bytes of the first column are used as OC(; .DH. The C bytes of OC(30OC(; .DH are both represented with K/, 3, C2, J/, #2, H;, #3, R3 and 7/. 2. High order channel overhead function +/, OC(30OC(; .DH channel trace byte% K/ This byte is used to repeatedly send high order channel access point identifiers +HD &.Id, so as to enable the receiving end of the channel to confirm that it and the specified transmitting end are in continuous connection status. The byte is used for the tracing of channel connection status. The K/ byte enables operators to find and eliminate faults in advance, prevent any transmitted service from being influenced and shortening the network restoration time. +2, Channel I.(< code% 3 The 3 byte +< bits, is used for the monitoring of channel error codes and is I. code using even parity check. I.(< code calculates all the bits of the first OC(30OC(; and the results are located at the 3 byte of current OC(30OC(;. +3, Signal label byte% C2 The C2 byte is used to stand for the composition or maintenance status of OC(30OC(; and the he$adecimal code corresponding to this byte and its meaning are shown in Table /(/. Table /(/ C2 byte code specification
<(bit code of C2 99999999 9999999/ 999999/9 9 9 9 9 9 9 // 99999/99 999/99/9 9 9 9 / 9 9 // 999/9/99 999/9/9/ /////// 9 //////// He$adecimal code 99 9/ 92 93 9; /2 /3 /; /= ### *eaning 8nloaded signals or monitored unloaded signals Indefinite payload loaded T8J structure 4ocked T8 3;.3><*bit0s and ;;.B3>*bit0s signals mapped to C(3 in asynchronous mode /3C.2>;*bit0s signals mapped to C(; in asynchronous mode &T* mapping *&7 +D5D , mapping #DDI 9. /</ test signal mapping OC(&IS +for cascade only,

+;, Channel status byte% J/ The J/ byte is used to send back to OC(30OC(; path source end the channel termination status and performance condition detected at the path sink end and thus enables any end of or any point of the path to monitor the status and performance of a full duple$ path. +=, Channel user channel bytes% #2 and #3 These two bytes are used for communication between channel units and are related to payload. +>, 4ocation indication byte% H; This byte provides general location indication for payload and can also indicate special payload locations, for e$ample, indication of the multi(frame location of OC(/0OC(2. +B, &.SN channel% R3 +b/:b;, These several bits are used as the &.S instruction of high order path(level protection. +<, 7etwork operator byte% 7/ This byte provides Tandem Connection *onitor +TC*, function for a high order channel.

+C, Spare bits% R3 +b=:b<, These several bits are reserved for future uses and their uses have not yet been specified.

1...) -rrangement of low order c$annel over$ead

/. 4ocation of a low order channel overhead byte OC(/0OC(2 .DH is made up of the bytes O=, K2, 72 and R;, which are respectively located at the first bytes +byte, of ; consecutive OC(/0OC(2 frames, i.e., every four OC(/0OC(2 .DH frames +=99As, are completely transmitted once. 2. 4ow order channel overhead function +/, O= byte The O= byte can provide the information about error code check, signal label and channel status of a OC(/0OC(2 channel. The first and second bits of the O= byte are responsible for channel error code performance monitor. The third bit and the fourth bit are used respectively for channel 6-I and for channel 6emote #ailure Indication +6#I,. The fifth, si$th and seventh bits provide OC( /0OC(2 signal label function and the eighth bit is used for OC(/0OC(2 channel 6emote Defect Indication +6DI,. +2, Channel trace byte% K2 This byte This byte is used to repeatedly send low order channel access point identifiers +4D &.Id, so as to enable the channel receive terminal to confirm that it and the specified transmitting end are in continuous connection status. +3, 7etwork operator byte% 72 This byte provides the TC* function for a low order channel. +;, &.S channel% R; +b/:b;, These several bits are used as the &.S instruction of low order channel(level protection. +=, 6eserved bit% R; +b=:bB, These several bits are reserved optional ones, whose uses depend on the path source end, which generates R; bytes. +>, Spare bit% R; +b<, This bit is reserved for future uses and its uses have not yet been specified.

1./ 0ogic composition of &D* e1uipment

SDH transmission network is made up of different types of 7-s connected via optical cable lines and performs the transmission function of an SDH network via different 7-s. These functions are add0drop services, cross(connection services, network fault self(healing, etc. &mong the commonly seen 7-s in an SDH network are Terminal *ultiple$er +T*,, &dd(drop *ultiple$er +&D*,, 6egenerator +6-J, and Digital Cross(connection System +DEC,.

1./.1 2erminal Multiple#er +2M,

& T* is used at a network terminal node, as shown in #ig. /(/9.

TM
140Mbit/s 34Mbit/s STM-M 2Mbit/s

STM-N

#ig. /(/9 Schematic diagram of model of a T*

The function of a T* is to multiple$ the low(speed signals of a tributary port to high(speed signal ST*(7 of a line port or to de(multiple$ low(speed tributary signals from ST*(7 signals. /( channel ST*(7 signals are input0output to its line port while multi(channel low(speed tributary signals can be output0input at a tributary port. 'hen low(speed tributary signals are multiple$ed into the ST*(7 frame of line signals, the locations of tributary signals in the line signals ST*(7 can be specified arbitrarily.

1./.2 -dd3Drop Multiple#er +-DM,

&D* is used at the transfer site of an SDH transmission network, such as the middle node of a link or a node in a ring, and is the most important 7- used most fre"uently in an SDH network, as shown in #ig. /(//.

STM-N

ADM
140Mbit/s 34Mbit/s STM-M 2Mbit/s

STM-N

#ig. /(// Schematic diagram of model of an &D*

&D* has two line ports and one tributary port. The two line ports are connected with optical cables on their respective sides +two transceiving optical fibers on each side,. #or the sake of description, we specify them as the 'est +', line port and -ast +-, line port. The function of &D* is to multiple$ low(speed tributary signals to lines +line singles, in cross(connection mode or de(multiple$ low(speed tributary signals from the line signals received from line ports. In addition, cross(connection of the ST*(7 signals on -astward0 westward line sides can be implemented. &D* is the most important 7- in an SDH network and can be e"uivalent to other 7-s, i.e., it can perform the functions of other 7-s. #or e$ample, &D* may be e"uivalent to two T*s.

1./.3 4egenerator
There are two kinds of regenerators in an optical transmission network. Dne is the pure optical regenerator, mainly used to amplify optical power so as to e$tend the optical transmission distance. The other is an electric regenerator used for pulse regeneration shaping and it can achieve the goal of accumulating no line noise and ensuring complete waveforms of transmission signals by means of Dptical0electric +D0-, conversion, sampling of electric signals, decision, regeneration shaping, -lectric0optical and other processing. Described here is the latter one, which

has only two line ports, as shown in #ig. /(/2.

STM-N

REG

STM-N

#ig. /(/2 Schematic diagram of model of a regenerator

The function of 6-J is to send the received optical signals from the offside after D0-, sampling, decision, regeneration shaping and -0D. &n 6-J in a real sense only needs to process 6SDH in the ST*(7 frame and needs no cross( connection function. However, &D* and T* need to process both 6SDH and *SDH because they are to insert low(speed tributary signals into ST*(7. In addition, both &D* and T* have the cross(connection function.

1./.) Digital Cross3connection &"stem +D5C,

The DEC is mainly responsible for the cross(connection of ST*(7 signals and is actually e"uivalent to a cross(connect matri$, which implements the cross(connection of various signals, as shown in #ig. /(/3.

M channel

DXC

N channel

#ig. /(/3 Schematic diagram of model of DEC

DEC can implement cross(connection of the input *(channel ST*(7 signals to the output 7( channel ST*(7 signals. The core of DEC is a cross(connect matri$ and the powerful DEC can implement the low priority cross(connection of high(speed signals in a cross(connect matri$. 8sually, DEC m0n is used to represent the type and performance of a DEC +m n,. m represents the ma$imum rate level which can be accessed to DEC and n does the minimum rate level of a cross(connection which can be implemented in a cross(connect matri$. The greater m is, the higher bearer capacity a DEC has. The smaller n is, the more fle$ible cross(connection DEC has. The meanings of corresponding values of m and n are shown in Table /(2. Table /(2 DEC m0n value rate correspondence table
m or n Correspond ent rate 9 >; kbit0s / 2*bit0s1 2 2*bit0s1 3 2*bit0s1 ; 2*bit0s1 2*bit0s1 = 2*bit0s1 > 2. =Jbit0s

1.6 P$"sical topolog" of an &D* transmission network

7etwork physical topology generally refers to the shape of a network, namely, geometric arrangement of network nodes and transmission lines, and reflects physical connectivity of

network nodes. The effectiveness, reliability and economy of a network depend to a great e$tent on the specific network architecture. There are = simple kinds of network physical topology structures, as shown in #ig. /(/;. +a, 4ine topology If all nodes in a communication network are cascaded with the first and last nodes open, a line topology is formed. In this topology structure, all nodes between two non(ad)acent nodes should be connected so as to implement the connection between non(ad)acent nodes. 4ine topology is an economical form of network topology used in an early SDH. +b, Star topology +)unction, If a special node is connected with all the other nodes, and there is no direct connection between them, a star topology is formed. In this topology structure, any two nodes e$cept the )unction nodes are connected with each other via a )unction node, which implements the routing and connection function for passing information stream. In this network topology, the nodes at a )unction center station can connect multiple optical fiber terminals to form a unified network, which implements integrated bandwidth management. +c, Tree topology If the ending node of a point(to(point topology unit is connected with several special nodes, a tree topology is formed. Tree topology can be considered as the combination of line topology and star topology. This topology structure is applicable to broadcast services, but not to bi(directional communication services because there e$ist such problems as bottleneck and restriction on optical power budget. +d, 6ing topology If all the nodes in a communication network are cascaded with no node open, a ring network is formed. If the first and last open nodes of a line network are connected, a ring network is formed. In a ring network, To connect two nodes, all the nodes between them should be connected. The best advantage of this network topology is its strong survivability, which is of vital importance to a modern high(capacity optical fiber network. Therefore, special importance is attached to a ring network in an SDH network. +e, *esh topology If many nodes in a communication network are directly interconnected, a mesh topology is formed. If all the nodes are directly interconnected, such mesh topology is called an ideal one. In a non(ideal mesh topology, any two nodes not connected directly with each other can be connected via the connection function of other nodes. The mesh structure is not influenced by the problems of node bottleneck and failure and there are multiple optional routes between two nodes. It has a high reliability, but a comple$ structure and high costs. Therefore, it is applicable to a backbone network with heavy traffic. In a word, all these topology structures have their own features and can be applied in a network to different degrees. 'hat network topology to be selected depends on many factors. #or instance, a network should be of high survivability and easy to configure, and the net architecture should be suitable for the introducing of new services. The different parts of an actual network are applicable to different topology structures. #or instance, ring topology and star topology structures are very applicable to a local network +namely, an access network or user network,, with a line topology structure sometimes used. 6ing topology and line topology are very applicable to a local interchange relay network while a toll network may demand mesh topology.

(a) !ine t"#"l"$%

TM

ADM

ADM

TM

(b) Sta& t"#"l"$%

TM

DXC/ADM

TM

TM

TM

TM

(c) T&ee t"#"l"$%

DXC/ADM

ADM TM TM

ADM

TM TM

(') Rin$ t"#"l"$% ADM

ADM ADM ADM

(e) Mesh t"#"l"$%

DXC/ADM

DXC/ADM

DXC/ADM

DXC/ADM

#ig. /(/; .hysical topology of an SDH network

1.7 &D* self3$ealing network

1.7.1 8etwork surviva(ilit"
*odern society relies more and more on communication and the survivability of a communication network has become a design inde$ of vital importance. The so(called self(healing network means that a network can automatically recover the carried services from a failure fault in a very short period of time without making users be aware of any network fault. Its basic principles are to enable a network to find faults and reestablish communication. & self(healing network involves no repairing and replacement of a specific faulty component or part but the reestablishment of communication. The former case still demands manual interference.

1.7.2 2"pes and principles of a self3$ealing network

&ccording to the definition of a self(healing network, there are many means to reali!e a self( healing network, but there are some common factors to be taken into consideration for various

self(healing networks% initial cost, proportion of traffic to be restored, additional capacity necessary for restoration of tasks, service restoration speed, fle$ibility in upgrade or addition of nodes, the feature of being easy to operate, run and maintain, etc. There are two basic forms of implementation of a self(healing network, namely, the line protection switching and self(healing ring network, which also has many types. Introduced below are the protection modes and implementation methods of various self(healing networks. /. 4ine protection switching The simplest self(healing network is the line protection mode commonly adopted in a conventional .DH system, which is also applicable to an SDH system. Its operating principle is that the system includes active0standby optical fibers. 'hen the service transmission of some operating fibers is interrupted or the performance is deteriorated to some degree, the system switching e"uipment will automatically switch active signals to the standby optical fiber for transmission. This protection mode has very short service restoration time, which may be less than =9ms, and is highly effective to any optical0electric component failure fault of a network node. However, if an optical cable is cut off, the active and standby optical fiber in the same cable core are both cut off. In this case, the above protection mode will not function any more. 'ith respect to the above(mentioned, the improvement of line protection switching can be made by means of geographically standby route, that is, the active and standby optical fiber can be laid respectively by means of different geographic routes. Thus, when the optical cable of an active channel is cut off, that of a standby path still will send signals safely to a remote end without being influenced. This standby route method is easy to configure and has simple network management. esides, it enables services to be restored very fast. However, this method needs at least double optical fiber cables and line e"uipment and a standby route is always very long, hence costs highly . In addition, this protection method can only protect transmission links and cannot provide the failure protection of a network node. Therefore it is mainly applicable to the protection of point( to(point applications. In such a case where there is stable and heavy traffic between two points, the standby route line protection method is still a very good protection means. 2. Self(healing ring network 7etwork nodes can be connected as a ring to further improve the survivability and cost of a network. &nd a ring self(healing network is also called a self(healing ring network. The node of a self(healing network can be either a DEC or an &D*, which is commonly used. & self(healing network employing the intelligent add0drop capability of an &D&* is one feature of SDH and also a field where active research is currently done. &ccording to different bases of self(healing ring service protection, a self(healing network can be categori!ed as path protection switched ring and multiple$ section protection switched ring. In terms of functional structure, the path protection switched ring is also called subnet connection protection and the multiple$ section protection switched ring is also called path protection. #or a path protection switched ring, the protection of service information is based on each path, which determines whether to switch based on the "uality of each path signal in a ring. #or a multiple$ section protection switched ring, the traffic protection is based on a multiple$ section, which determines whether to switch based on the "uality of multiple$ section signals between each pair of nodes. 'hen any fault occurs to a multiple$ section, all multiple$ section service signals between nodes are switched to a protection loop. Dne important difference between a path protection switched ring and a multiple$ section protection switched ring is that the former always

uses private protection. That is, a protection section transmits service signals in normal cases. However, the latter always uses shared protection, that is, a protection section is idle in normal cases and a protection time slot is shared by each pair of nodes. In terms of the direction in which information between nodes in a ring is transmitted, a self( healing ring can be categori!ed as unidirectional ring and bi(directional ring. In normal cases, all received0transmitted service signals in a unidirectional ring are transmitted in the same direction +clockwise or anticlockwise, in a ring and those in a bi(directional ring are transmitted in the reverse direction in a ring. &ccording to the applications of service paths and protection paths, there e$ist such protection modes as /%/, /P/, etc. in a self(healing. /%/ protection mode means that in normal cases, service signals are only connected with an operating path and additional service signals are transmitted in a protection path. 'hen any fault occurs to an operating path, the nodes will abandon the additional services in a protection path and switch them into transmission service signals to achieve the protection of service signals. /P/ protection mode means that service signals )ump simultaneously between an operating path and protection path, and the nodes to receive services receive preferential service signals from them. That is, when any fault occurs to an operating path, the nodes switch automatically to a protection path and receive service signals. &nd vice versa. In terms of the minimum number of optical fiber used between each pair of nodes in a ring, a self( healing ring can be categori!ed as two(fiber ring and four(fiber ring. ased on the above various classification methods, we can obtain multiple different self(healing ring structures. In normal cases, the path protection switched ring works in the unidirectional two( fiber mode. The multiple$ section protection switched ring can be in either unidirectional mode or bi(directional mode. &nd it can be either two(fiber mode or four(fiber mode. 7ow take a four(node ring for e$ample to introduce four typical and practical self(healing ring structures. /N 2(fiber unilateral channel changeover ring The two(fiber unidirectional path protection switched ring is shown in #ig. /(/=.
C& &C S/
./

C& &C S/ ./ & D ./ S/ C& &C S(itchin$

& D
./ S/

C& &C

#ig. /(/= Two(fiber unidirectional path protection switched ring

The protection mode of the 2(fiber uni(directional channel changeover ring is the same as the channel /P/ protection, also based on the Gsend together and receive firstH principle, with reference to .&TH(&IS, with no need for an &.S protocol. It has two pieces of optical fiber, one being S optical fiber used for transmission of service signals and the other being . optical fiber used for protection. The two(fiber unidirectional path protection switched ring adopts the structure Gbridge connection for the initial end and switching for the last endH. That is, in the nodes & and

C, the signals entering a ring are accessed to S optical fiber and . optical fiber simultaneously while the signals of branch nodes are obtained by means of switching. &s shown in #ig. /(/=, in node &, the tributary signals entering a ring and reaching the destination node C are simultaneously fed in the transmitting optical fiber S/ and ./. S/ optical fiber sends service signals to the branch node C in the clockwise direction while ./ optical fiber sends identical tributary signals to the branch node C in the reverse direction. 7ode C receives the signals from two directions at the same time, and it will choose one as the tributary signals according to their "uality. 7ormally, the signals sent from S/ are the predominant signals. &ccording to the direction of the tributary signals entering a ring and that of the tributary signals returning from this tributary signal branch node, a ring can be categori!ed as the unidirectional ring and bi(directional ring. In #ig. /(/=, the signals from & to C and those returning from C to & are all transmitted along S/ in clockwise direction and along ./ in anticlockwise direction, therefore it is a unidirectional ring. 'hen the optical cable between and C nodes is cut off, the switching switch in the node C will turn from S/ to ./ on the principle of transmitting simultaneously and receiving preferentially and receive the &C signals from the node & via ./ as the branch signals because the &C signals coming from & via are lost. Thus, the service signals between & and C are maintained instead of being lost. 'hen the faults are removed, the switches return to their former states. 2, Two(fiber uni(directional multiple$ section switched ring The two(fiber uni(directional multiple$ section switched ring is shown in #ig. /(/>.
C& &C S/ ./ & D ./ S/ D C & C& &C

S . / /

. S / /

C& &C

C& &C

Switching

#ig. /(/> Two(fiber uni(directional multiple$ section switched ring

&s shown in #ig. /(/>, in a two(fiber unidirectional multiple$ section switched ring, the node has a protection switched switch in each high(speed line before tributary signal add0drop. In normal cases, low(speed tributary signals can be added0dropped only via S/, with ./ idle. The signals from & to C and those returning from C to & are all transmitted along S/ in clockwise direction, therefore it is a unidirectional ring. 'hen the optical cable between and C nodes is cut off, the protection switched switches in them will e$ecute the loopback function via &.S protocol. In the node , the &C signals on S/ return from ./ via the switched switch and reach the node C via the nodes & and D in anticlockwise direction. Then, they loop back to S/ via the switched switch of the node C and branch out. This loopback switching function can ensure that the continuity of a ring is still maintained in faulty status and service signals on low(speed tributaries will not be interrupted. 'hen the faults are removed, the switches return to their former states.

3N #our(fiber bi(directional multiple$ section switched ring The four(fiber bi(directional multiple$ section switched ring is shown in #ig. /(/B.
C& &C
S/ ./ S2 .2

C& &C
S/ ./ S2 .2

& D
.2 S2 ./ S/

& D
.2 S2 ./ S/

C& &C

C& &C

Switching

#ig. /(/B #our(fiber bi(directional multiple$ section switched ring

&s shown in #ig. /(/B, the four(fiber bi(directional multiple$ section protection switched ring has two pieces of service optical fiber S/ and S2 and two pieces of protection optical fiber ./ and .2. The four pieces correspond respectively to receiving0transmitting directions. In normal cases, the low(speed tributary signals entering a ring from the node & and reaching the destination node C are transmitted along S/ in clockwise direction. However, the low(speed tributary signals returning from the node C to the node & are transmitted along S2 in anticlockwise direction, therefore it is a bi(directional ring. The protection optical fiber ./ and .2 are both idle. 'hen the optical cable between and C nodes is cut off, the two respective protection switched switches in them will e$ecute the loopback function via &.S protocol to maintain the continuity of a ring. The optical fiber S/ communicates with the optical fiber ./ while S2 communicates with .2. The &C signals along S/ return from ./ via the switched switch of the node and reach the node C via the nodes & and D in anticlockwise direction. Then, they return to S/ optical fiber via a switched switch and branch out. The C& signals are similar to this. It can be seen from this figure. that the service signals on S/ and the protection signals on .2 are transmitted in completely the same direction, namely, in clockwise direction. y means of timeslot interchange technology, the signals on the optical fiber S/ and .2 can be made to be in the same optical fiber, which is called S/0.2 optical fiber. In this case, half of the timeslots on this optical fiber such as odd time slots are used for transmission of service signals while the other half such as even timeslots are for transmission of protection signals. 4ikewise, there is S20./ optical fiber. The protection signal timeslots on S/0.2 can protect the service signals on S20./ while those on S20./ can protect those on S/0.2. Thus, a four(fiber ring can be simplified as two(fiber ring. #or a two(fiber bi(directional multiple$ section switched ring, we usually adopt odd0even timeslot protection. Df course, there are other protection modes, for e$ample, the first half of timeslots transmit service signals while the second half do protection signals. ;N Two(fiber uni(directional multiple$ section switched ring The two(fiber bi(directional multiple$ section switched ring is shown in #ig. /(/<.

C& &C S/0.2 S20./ & D S20./ S/0.2 C D S20./ S/0.2

C& &C S/0.2 S20./ & C

C& &C

C& &C

Switching

#ig. /(/< Two(fiber bidirectional multiple$ section switched ring

&s shown in #ig./(/<, the &C signals are transmitted along S/0.2 optical fiber in clockwise direction while the C& signals along S20./ in anticlockwise direction. Therefore, it is still a bi( directional ring. 'hen the optical cable between and C nodes is cut off, the switched switches in them will communicate S/0.2 and S20./ according to &.S protocol. y means of the timeslot interchange technology, the service signals timeslots on S/0.2 and S20./ can be shifted to the protection signal timeslots of another piece of optical fiber to implement the protection switching function. The time for the protection switching is less than 39ms. #or e$ample, the service signal odd timeslots on S/0.2 can be shifted to the protection signal even timeslots on S20./. That is, all service signals are transmitted in a piece of optical fiber and the timeslot interchange like this is needed in the four sites &, , C and D. 'hen the faults are removed, the switches return to their former states.

1.9 2iming s"nc$roni%ation

7etwork synchroni!ation is one important part in network planning, especially so for an SDH network on the basis of synchronous transmission. Dnly when network synchroni!ation is reasonably planned can the optimal synchroni!ation effects be achieved between 7-s. SES*(/=9 +O2, e"uipment provides the synchroni!ation plane based on SS* information. Synchroni!ation Status *essaging +SS*, can be used to ensure that an 7- selects an effective timing source of the best synchroni!ation "uality, prevent timing from forming loop and guarantee the timing synchroni!ation performance of a network.

1.9.1 &&M function of an &D* interface

The service add0drop and rerouting capabilities of an SDH network enable a network to be applied with unprecedented fle$ibility and high survivability and makes selection of network synchroni!ation timing more comple$. In an SDH network, the timing reference allocation between nodes is made by means of a great number of low(level SDH network clocks, therefore the "uality of the timing reference must be labeled by some means. SS* is right used to display the information of the timing reference "uality. SS* is transferred by the =th:<th bits of S/ byte in an SDH multiple$ section overhead, as shown in #ig. /(/C.

b1

b2

b3

b4

b(

b%

b&

SS*
#ig. /(/C Contents of S/ byte

These four bits have /> different kinds of codes representing /> different synchroni!ation "uality grades, as shown in Table /(3. Table/(3 SS* code Descriptions of SDH synchroni!ation "uality grades 9999 8nknown synchroni!ation "uality +e$isting synchroni!ation network, 999/ 6eserved 99/9 J. <// clock signal 99// 6eserved 9/99 J. </2 transit e$change clock signal 9/9/ 6eserved 9//9 6eserved 9/// 6eserved /999 J. </2 local e$change clock signal /99/ 6eserved /9/9 6eserved /9// Synchronous -"uipment Timing Source +S-TS, //99 6eserved //9/ 6eserved ///9 6eserved //// 7ot to be used as synchroni!ation In an SDH network, the timing reference allocation between nodes is made by means of a great number of low(level SDH network clocks. 'ith the increase in 7-s on the synchroni!ation links, the "uality of timing reference signals degrade gradually. Therefore, when there are multiple optional synchroni!ation paths of the same "uality grade in an 7-, selection of the synchroni!ation path passing the smallest number of 7-s helps improve the timing performance of an SDH network. Dn this principle, ST- designs S/ byte patent algorithm to enable an 7- to select the clock reference signals of the highest "uality grade and with the shortest synchroni!ation path. In an SDH network, the selection of clock source of SDH 7- is made mainly based on S/ byte and the following principles should be followed% /. 'hen there are multiple optional effective clock sources in an 7-, the 7- first selects the clock of the highest "uality grade based on the "uality grade information of clock source. If the clock sources are of the same "uality grade, the 7- will select the one passing the smallest number of 7-s based on the number of 7-s a clock source transmission path passes. 2. The 7- transfers via S/ byte the "uality grade information of the currently adopted clock source and the number of 7-s it passes to a downlink 7- and sends the status information G8navailableH to an uplink 7-. +&n uplink 7- is relative to a downlink one. If the 7- e$tracts a clock from the 7- &, the 7- & is the uplink 7- of the 7- and the 7- is a downlink one as S/ +b=:b<,

opposed to the 7- & #ig. /(29 is an application e$ample of SS*.

.6C
.6C

.6C 8available

& D
.6C

8available

C
.6C 8available

.6C

Synchroni!ation path +In use, Synchroni!ation path +7ot in use,

.6C 8available

SS*

+a, +,C
+, C

+, C
)na*ailable

A *
+,C

)na*ailable

) C
)na*ailable +,C

+,C S%nch&"ni+ati"n #ath (,n -se) S%nch&"ni+ati"n #ath (N"t in -se) .RC N"t in -se

SSM

( - b- )

SETS

SETS )na*ailable

A *
SETS

)na*ailable

) C
SETS SETS

)na*ailable S%nch&"ni+ati"n #ath (,n -se) S%nch&"ni+ati"n #ath (N"t in -se) SETS )na*ailable

SSM

( - c- )

#ig. /(29 &pplication of SS*

&s shown in #ig. /(29, the 7- & is accessed to an e$ternal clock source +.6C, in the transmission network shown in +a, and the synchronous source settings of various 7-s are shown in Table /(;. Table/(; Setting of 7- synchronous source 7etwork -lement Clock source list 7- & -$ternal clock source and internal clock source 74ine clocks / and 2 7- C 4ine clocks / and 2 7- D 4ine clocks / and 2 The 7- & transfers the "uality grade information +.6C, of e$ternal synchroni!ation to other 7-s. In the 7- , there are optional synchronous sources &( line clock and &(D(C( line clock. Dn the principle of the minimum 7- a timing path passes, the 7- will automatically select &( line clock as its synchronous source. Similarly, the 7- D will automatically select &(D line clock. The 7- C can select either &( (C line clock or &(D(C line clock. Oarious 7-s will send the status information G8navailableH to their uplink 7-s. 'hen the line between the 7- and the 7- C is blocked, the 7- C will select &(D(C line clock as shown in #ig. /(29 +b,. &s shown in #ig. /(29 +c,, when the e$ternal clock source connected to the 7- & is interrupted, the 7- & enters the clock holdover mode, after which the mode will be free(run mode. In this case, various 7-s are still synchronous with the 7- & and the clock source grade degrades as the Synchronous -"uipment Timing Source +S-TS, of the 7-.

1.9.2 &&M function of a s"nc$ronous interface

The system can provide a synchronous interface of 2*H! or 2*bit0s, which supports SS* function and is used for SS* of .DH. SS* information is transmitted via one +what bit to be used depends on the user, of idle bits S an +n being ;, =, >, B or <, and representing the ; th:<th bits in TS9

timeslot, in TS9 timeslot of odd frames in J. B9; multiframe. #our odd frames are considered as a group and one San bit in each odd frame forms four and a half bit bytes from S an/ to San;, as shown in #ig. /(2/. Then, they are used to define synchroni!ation "uality grades and both their bit patterns and synchroni!ation "uality grades are defined as S/ byte in SDH is +please refer to Table /(3,.

San/&San2&San3'San;
n@;&=&>&B&<
#ig. /(2/ Diagram of Sa bit organi!ation

1.1 2ransmission damage

1.1 .1 :rror code c$aracteristics
/. Definition of error code So(called error code is that errors occur to some bits of a data stream after being received and regenerated and the "uality of transmission information is damaged. Jenerally, long(term mean it -rror 6ate + -6, is used to evaluate the "uality of information transmission, that is, the ratio between the number of error bits and the total number of transmitted bits within a specific period of observation time is considered as -6. How error codes influence various services depends mainly on the kinds of services and the distribution of error codes. #or e$ample, in voice communication, the effect of random error code is only the click of the receiver, whose influence on conversation "uality can generally be tolerated. However, information in data communication almost has no redundancy and the whole block will be out(of(service if errors occur to one bit in a block. esides, errors occurring to a bit or multiple bit strings in a block may produce the same effects. Therefore, it can be considered that voice communication can tolerate of random error code distribution while data communication can tolerate bursty error code distribution. +/, -rror code occurring internally -rror codes of this kind in the system includes those produced by various noise sources, those resulting from location )itter, those produced by a multiple$er, cross(connection e"uipment and e$change and those resulting from intersymbol interference produced by optical fiber dispersion. -rror codes of this kind can be shown by long(time system error code performance. +2, -rror codes resulting from pulse interference -rror codes caused by bursty pulses such as electromagnetic interference, e"uipment fault, transient power interference, etc. -rror codes of this kind are of bursty nature and occur on a large scale. 'hen a great number of error codes occur in the system, the short(term error code performance of the system will show this. 3. -rror code performance specifications Currently, IT8(T specifies 3 high(bit(rate channel error code performance parameters% +/, -rrored Second 6ate +-S6, 'hen there is one or multiple faulty blocks within / second, this second is called an -rrored Second +-S,. The ratio between the number of -Ss with a specified period of measuring time and

the total time available is called -S6. +2, Severity -rrored Second 6ate +S-S6, 'hen there are faulty blocks no less than TM +T temporarily specified as 39, or at least one Severity Disturbed .eriod +SD., within / second, this second is considered as Severity -rrored Second +S-S,. The ratio between the number of S-s with a specified period of measuring time and the total time available is called S-S6. S-S is generally the bursty error block resulting from pulse interference, therefore S-S6 always represents the antiinterference capability of the e"uipment. SD. means that during the measurement, the error code rates of all consecutive blocks within the time segment at least e"uivalent to four consecutive blocks or /ms +taking the longer time segment of both, are more than or e"ual to /9(2 or signal loss occurs. +3, ackground lock -rror 6ate + -6, So(called - means the ones after the error blocks occurring during the unavailable time and S-S are deducted. The ratio between the number of -s and total number of blocks after all blocks occurring during the unavailable time and S-S are deducted is called -6. ecause big bursty error codes resulting in S-S and the unavailable time have been deducted during the calculation, the si!e of this parameter can generally represent the background error code of the system. -6 measured within a long period of time always represents the error code produced within the e"uipment and are related to the performance stability of the e"uipment components. The evaluation of error code performance parameters can be meaningful only when a channel is available. IT8(T specifies that the unavailable time starts with the time when /9 consecutive S-S events occur and ends with the time when /9 consecutive non(S-S events occur. Then, the available time begins with the starting moment of these /9 seconds.

1.1 .2 ;itter c$aracteristics

/. Definition and influence of )itter Timing )itter +)itter for short, is defined as the short(term deviation of a specific moment +for e$ample, the optimal sampling moment, of digital signals from their ideal time. The short(term deviation means the phase change with the change fre"uency higher than /9H! and the phase change with the change fre"uency lower than /9H! is called wander. Timing )itter damages the network performances in the following aspects% #or analog signals with digital coding, random phase )itter of decoded digital streams makes the recovered sample values have irregular phases, which results in the distortion of output analog signals to produce so(called )itter noise. In a regenerator, irregular timing makes effective )udgment points deviate from the center of receiving eye pattern, thus reducing the signal0noise ratio redundancy of the regenerator till error codes occur eventually. In an SDH network, for network units similar to the synchronous multiple$er e"uipped with a cache, too much input )itter will make a cache overflow or vacant, thus resulting in error codes. Kitter influences various services in different ways. The voice signals with digital coding can endure very great )itter and allow of root(mean(s"uare )itter of up to /. ;US. However, because human eyes are sensitive to phase changes, a color TO with digital coding can tolerate much less )itter. 2. Jeneration of )itter

+/, Kitter of line system Kitter of a line system can generally be categori!ed as random )itter and systematic )itter. The former is )itter component generated by various regenerators but unrelated to each other. The latter generally refers to amount of )itter related to each other and generated by various regenerators. It is always related to signal patterns, hence also called pattern(related )itter. ecause there is no line code in an optical synchroni!ation line system e$cept for simple scramble, its patter(related )itter may somewhat increase )ust to make systematic )itter accumulated at a greater speed. /, 6andom )itter source Oarious noise sources% in an optical cable system, there is avalanche noise, "uantum noise, kink noise, mode allocation noise and reflection noise besides common thermal noise. These noises are different in mechanism, but they will result in random distortion of signal pulse waveforms and random phase spurious modulation of the output signal waveforms of a timing filter. Thus, )itter occurs. Detuned timing filter% & detuned timing filter will produce asymmetric output waveforms e"uivalent to asymmetric sideband modulation and cause modulation of clock component amplitude and phase. .hase modulation will result in timing )itter. Completely irrelevant pattern )itter% .atter(related )itter is made up of two "uadrature components, namely, completely relevant pattern )itter and completely irrelevant pattern )itter. The latter is different from general random )itter in mechanism, but with the same law of accumulation, therefore it can also be included in this type. 2, Systematic )itter source In an ideal regenerator, signal pattern has no influence on the phase of an output timing signal. ut a regenerator actually will have many defects, which will result in phase change of timing signals and produce )itter. Intersymbol interference% To reduce the cost of an e"uali!er, some small amount of intersymbol interference is generally permitted. ut with the temperature changing and components ageing, intersymbol interference will increase and make signals produce random wander of output pulse peak location after passing nonlinear components. &nd timing )itter occurs. &ction of limited pulse width% Since the signal pulse width cannot be infinitely narrow, its spectrum has a section inclined continuum near the clock. Its tilt varies with the contents of transmitted information and this time(variation inclination will result in patter(related )itter of timing signals. Threshold wander of amplitude limiter% The threshold of an amplitude limiter will wander with the temperature changing and components ageing. 'hen signals with different amplitudes are added to an amplitude limiter, they will have different cross points with a threshold, which make output pulse locations vary with the amplitudes of input signals. &nd the amplitudes of input signals are related to the pattern of transmitted information, thus forming pattern(related )itter. .attern effect of a laser% In a high(bit(rate system, because the pulse repeated cycle becomes shorter, the limited on0off time of a laser will have more influence on the patterns of transmitted signals. This influence will vary with different patterns of transmitted signals, so that pattern( related )itter is caused. It can be seen from the above mentioned that the )itter source and its physical mechanism in an optical cable line system are very comple$. Currently, the random )itter root(mean(s"uare value of a typical regenerator is 9.=:/.=, the systematic )itter root(mean(s"uare value is 9.= :3.= and the

total )itter root(mean(s"uare value is about / :;. Jiven that 3>9 is a 8nit Interval +/8I,, the )itter root(mean(s"uare value of a typical regenerator is 9.9938I:9.9/98I. +2, *ultiple$er )itter The )itter mechanism of a multiple$er +including DEC, in an SDH network is "uite different from that of a multiple$er in a traditional .DH network. In a .DH multiple$er, main )itter comes from stuffing )itter introduced in )ustification. ecause the )ustification is made with the unit being bit, there is not much influence. In an SDH network, tributary signals are synchroni!ed by means of so(called pointer )ustification. That is, increase0decrease in pointer value is used to ad)ust the phase change and fre"uency change of low(speed tributary signals and this is conceptually very similar to traditional positive0!ero0negative )ustification. In a .DH network, the )itter fre"uency resulting from the plesiochronous multiple$ing process ranges from about some tens of Hert! to some one hundred Hert!. Therefore, when there is very low fre"uency component, the )itter and wander power are very small, which provides a relatively smooth channel for the tracing and control of the phase change of a reference main clock. However, pointer )ustification in an SDH network is made by the byte. That is, a byte includes < bits and byte )ustification at one time will generate <8I phase steps. #or /;9*bit0s tributary signals, pointer )ustification is made every three bytes and )ustification at one time will generate 2;8I phase steps. The digital signals with these phase steps will generate very long phase transition process when passing band limiting circuits. Dn the other hand, the pointer )ustification in an SDH network in the status of normal synchroni!ation is mainly caused by random noise in the course of synchronous allocation. Therefore the phase step resulting from pointer )ustification occurs irregularly, that is, the interval between any two phase steps has no upper limit and the whole phase )ustification will take a very long time. Therefore, the combination of pointer )ustification and network synchroni!ation will generate )itter or wander of very low fre"uency at SDH0 non(SDH boundary. The output )itter of a multiple$er is closely related to whether it is synchronous with the e"uipment and how it is. During normal synchroni!ation, the output )itter and wander depend on the filtering characteristics of an internal clock and the performance of synchronous timing reference. 'hat should be worthy of special notice is that in a cross(connection e"uipment and add0drop multiple$er, a high(fre"uency ST*(7 clock sometimes may be obtained from a low( fre"uency one after fre"uency doubling. In this case, the )itter on the original low(fre"uency clock will increase by the same times in linear mode, which is shown the output )itter of ST*(7, namely, so(called )itter multiplication. Kitter on a low(fre"uency clock is most probably to occur during active0standby clock switching. In this case, a phase transient process will occur, which results in clock timing )itter. &mong possible solutions are non(traumatic conversion, separate transmission of SDH synchroni!ation timing and strict filtering of a )itter clock. +3, .lesiochronous tributary output )itter of SDH0non(SDH boundary SDH is gradually developed from the environment of an original .DH network and at the )unction between them, the original .DH system has specified )itter clearly. Therefore, after SDH is introduced, the plesiochronous tributary output )itter of SDH0non(SDH boundary must also observe the rules concerned. ecause some new )itter mechanisms are introduced into SDH, special measures must be taken to satisfy the above indices. /, *apping )itter of plesiochronous tributary The mapping )itter of plesiochronous tributary means the )itter of terminal output signals when no )itter or pointer )ustification occurs to plesiochronous input signals.

SDH maps any plesiochronous tributary signals into ST*(/ frame structure and transmits them by means of stuffed bits. &t SDH gateway, any tributary signals will be recovered as long as stuffed bits and path overhead are removed. However, these tributary signals contain the gap resulting from the above bits being removed. To smooth the phases of these signals with gaps and reduce )itter, a cache and a phase smooth circuit are generally needed. In an SDH, they are called desynchroni!ers, which are actually implemented by the phase lock loop with caches. Df course, this desynchroni!er can also reduce other )itter resulting from the de(multiple$ing process. 2, .ointer )ustification )itter There are two working modes in an SDH network, namely, non(deterioration mode and deterioration mode. In the non(deterioration mode +namely, in normal operation,, all network element clocks are locked within a reference primary clock and no pointer )ustification will occur in an ideal case. However, the synchroni!ation allocation noise process is inevitable, therefore it is still possible that small amount of random pointer )ustification will occur. In the deterioration mode, when the source or terminal of an SDH network loses a timing reference, pointer )ustification results mainly from clock fre"uency shift. In whatever case, the process of pointer )ustification is represented une$ceptionally as plesionchronous tributary output )itter. ecause pointer )ustification is achieved by the single byte or three bytes while mapping is by the stuffed single bit, pointer )ustification will have much greater influence on phase )ump. &nd it is the main )itter source of SDH0non(SDH boundary and some measures must be taken to suppress it. 3. Kitter performance &s follows are the common parameters to measure )itter performance in an SDH network% /, Tolerance of input )itter Tolerance of input )itter includes that at the .DH input interface +tributary interface, and that at the ST*(7 input interface +line interface,. #or a .DH input interface, tolerance of input )itter means the ma$imum input )itter this input interface can withstand with no error code occurring to the e"uipment. To meet the needs for SDH 7-s to transmit .DH services in a transmission network, the tributary output interface of this SDH 7- must be able to contain the ma$imum )itter of .DH tributary signals. That is, the )itter tolerance of this tributary interface can withstand any )itter of .DH signals transmitted. The tolerance of input )itter at a line interface +ST*(7, is defined as sine peak(peak )itter value, which enables optical e"uipment to produce /d optical power penalty. This parameter is used to specify that the tolerance of input )itter of an 7- at this level should be able to contain the output )itter of an 7- at an upper level when SDH 7-s are interconnected to receive any ST*(7 signals. 2,. Dutput )itter Similar to the tolerance of input )itter, output )itter also includes that at the .DH tributary interface and that at the ST*(7 line interface. Dutput )itter is defined as the ma$imum )itter of output port signals when no )itter occurs to e"uipment input terminal signals. The output )itter of the .DH tributary port of SDH e"uipment should ensure that the )itter of output .DH signals should not e$ceed what the e"uipment receiving this signal can withstand when an SDH 7- transmits .DH services. The output )itter of the ST*(7 line port should not e$ceed what the remote SDH 7- receiving this ST*(7 signal can withstand. 3,. *apping and combined )itter ecause pointer )ustification and mapping at .DH0SDH network boundary will result in special SDH )itter, mapping )itter and combined )itter are used to describe this )itter and normali!e it.

The mapping )itter means the ma$imum )itter of output .DH tributary signals at the .DH tributary port of SDH e"uipment when .DH signals with different fre"uency shifts are input at the .DH tributary port of SDH e"uipment and no pointer )ustification occurs to ST*(7 signals. The combine )itter means that some pointer test se"uence signals conforming to J. B<3 specifications are input at the line port of SDH e"uipment. &nd pointer )ustification occurs to SDH e"uipment and the fre"uency shift of input signals will be somewhat changed. In this case, the ma$imum )itter of output signals measured at the .DH tributary port of the e"uipment is the combined )itter of the e"uipment. ;, Kitter transfer function(()itter transfer characteristics Kitter transfer function is defined as how the ratio between the )itter of e"uipment output ST*(7 signals and that of e"uipment input ST*(7 signals vary with )itter fre"uency. This characteristic aims at normali!ing how the )itter of e"uipment output ST*(7 signals suppress that of input ST*(7 signals +namely, )itter gains, to control )itter accumulation of a line system. ;. 'ander characteristics 'ander is defined as long(term wander of a specific moment +for e$ample, the optimal sampling moment, of digital signals from its ideal time. Here long(term wander means the phase change whose change fre"uency is lower than /9H!. 'ander causes both transmission signal bits and input signal bits to deviate from ideal time locations and makes input signal bits unable to be recogni!ed in a )udgment circuit. In this case, error code will occur. Dne way to reduce error code of this type is to add caches to the interface between transmission lines and terminal e"uipment to synchroni!e data one more time. The method is to write data into a cache by means of the clock e$tracted from receiving signals and then perform read operations of the cache with a same reference clock so as to forcedly synchroni!e various data streams of different phases. It is certain that in order not to overflow or be vacant, the capacity of the cache must be greater than the ma$imum possible input peak(peak wander, which is hard to achieve in practice. Therefore, in pro)ects, a cache is generally re"uired to contain any possible change in transmission delay within one day. However, those large(scale wanders with very low fre"uency are allowed to e$ceed the cache threshold, transform into slip damage and become part of uncontrolled slip indices allowed in a network. It can be seen that very small wander can be absorbed by a cache while those large(scale wanders will be transformed into slip in the end. The influence of slip on various services depends to a great e$tent on the rate of services themselves and information redundancy. The greater the speed is and the smaller information redundancy is, the greater influence slip has. Telephone services have great information redundancy, therefore slip has little influence on them. The influence of wander is similar to pulse noise caused by error code. Data signals almost have no redundancy, therefore they are seldom influenced by slip. #or e$ample, for packet data with a fi$ed length, slip will cause >;kbit0s path multi(frame to be out of frame. #or fa$ services without any error control, one slip may lose at most two scanning lines and automatically substitutes for it the previous two error(free scanning lines. Thus, definition is made slightly poorer. The most common cause for wander is the change of ambient temperature, which may lead to first the change in transmission characteristics of optical cables and then slow change of transmission signal delay. Therefore, wander can be understood simply as slow change of signal transmission delay and this transmission damage cannot be thoroughly solved by optical cable system itself. In

an optical synchroni!ation line system, there is still another type of wander caused by the combination of pointer )ustification and network synchroni!ation and this type can be reduced with some additional measures taken.

1.11 Basic knowledge of 2elecom Management 8etwork +2M8,

1.11.1 Principles of 2M8
/. T*7 management framework To reali!e unified and highly efficient management on telecom networks, IT8(T has raised the concepts of the telecom management network +T*7,. T*7 is independent of a telecom network and implements special network management. T*7 provides an organi!ational network architecture by means of a unified architecture with a series of standard interfaces +including protocols and message regulations, and enables various 7*Ss to interconnect with Telecom e"uipment. Thus, automatic and standardi!ed management of a Telecom network can be achieved, with various management functions provided as well. T*7 always employs partial facilities of a Telecom network to provide communication, therefore the two may be partially overlaid. #ig. /(22 describes the relation between a T*7 and a Telecom network.

TMN
/#e&atin$ s%ste0 /#e&atin$ s%ste0 /#e&atin$ s%ste0

Data c"00-nicati"n net("&1

3"&1 statin

E2chan$e

T&ans0issi"n s%ste0

E2chan$e Telec"0 net("&1

T&ans0issi"n s%ste0

E2chan$e

#ig. /(22 T*7 and Telecom network

2, T*7 physical structure T*7 physical structure mainly describes the physical entities and interfaces inside T*7, the simplified T*7 physical structure as shown in the diagram. DS in #ig. /(23 stands for an operating system, namely, 7*S. It is the system to e$ecute DS# and is actually a large(scale system program to manage network resources. *D stands for coordination e"uipment and e$ecutes *#. *D mainly coordinates DS and 7-, and can also provide 5&# and 'S#, sometimes even DS#. *D can be implemented in hierarchical mode. 5& stands for 5 adaptor and is responsible for the adaptation and interconnection between an 7- and a non(T*7 interface. The data communication network DC7 is a communication network in T*7 that supports DC#. It mainly reali!es the 3 lower layer functions of the DSI reference model, but not providing the ; th to Bth layer functions. DC7 can be formed by the interconnection of different types of sub(networks +such as E. 2= or DCC,. &n 7- is made up of the telecom e"uipment +or some of it, to e$ecute 7-# and support e"uipment. &n 7- may include other T*7 blocks, *# being the most common. It can contain

other T*7 function modules, and most often, *#. 7ormally, 7- has one or more standard 5 interfaces, and also the # interface. The workstation 'S is the device that e$ecutes 'S#, mainly performing the conversion functions between the f reference point information and the g reference point displaying format.
T*7 DS 530#0E DC7 53 530# 53 *D 5$ DC7 5$ 5$ 5& 75& 7'S

#ig. /(23 Diagram of T*7 physical structure

3. T*7 interface To simplify the interconnection between e"uipment of multiple manufacturers, the standard T*7 interface shall be stipulated, which is one of the keys to T*7. The standard interface re"uires uniform specifications about the protocol stacks, and the messages carried by protocols. 5 interface% Dften, the 5 interface corresponds to the 5$ and 53 interfaces. &mong them, the 5$ interface connects *D and *D, 7- and *D, 5& and *D, and 7- and 7- +at least one performs the *# functions,. ut the 53 interface connects *D, 5&, 7- and DS to DS via DC7. In the traditional .DH system, the 5$ interface usually performs the 3 lower layer functions of the DSI reference model, so it is suitable to connect such simple devices as the multiple$er and line systems. Its protocol stack can select the &/ or &2 protocol stacks in the IT8(T recommendation J. BB3. The former is the connection(oriented mode, while the latter works in connectionless mode +4&7 technology,. In the SDH system, the 5$ interface usually contains the functions of the whole B layers. Its protocol stack can select the CD7S/, C47S2 and C47S/ in the IT8(T recommendations 5. <// and 5. </2. Here, CD7S/ is the E. 2= packet networking interface, C47S/ is the connectionless mode interface using the 4&7 technology, but C47S2 is the connectionless mode interface using the interconnection protocol on basis of the E. 2= protocol. The 53 interface has the functions of all the B layers, suitable for comple$ e"uipment such as switches and DEC. Its protocol stack is also selected from 5. <// and 5. </2. # interface% # interface corresponds to the f reference point and can connect a remote work station to an DS or *D via DC7. The J interface corresponds to the g reference point and the E interface to the $ reference point. 8sually, the E interface has higher re"uirements for safety than the 5 interface. ;. Classification of T*7 layers The management layer model of T*7 can be classified according to IT8(T *. 39/9 as% 7etwork -lement 4ayer +7-4,, network -lement *anager 4ayer +-*4,, 7etwork *anagement 4ayer +7*4,, Service *anagement 4ayer +S*4, and usiness *anagement 4ayer + *4,. #ig./(2; describes the classification of T*7 management layers. In this figure, 7- can be either SDH e"uipment or any manageable e"uipment like .DH or an e$change.

Se&*ice la%e& 0ana$e0ent s%ste0

SMS

NMS
NE! 0ana$e0ent s%ste0

NMS

N! 0ana$e0ent s%ste0

EMS

EMS

NE! NE NE NE

NE NE

NE

NE

#ig. /( 2; T*7 management layers

1.11.2 &D* Management 8etwork +&M8,

/. S*7 and T*7 The SDH management network +S*7, is a sub(set of T*7 that manages SDH network elements. It can be further divided into a series of SDH management sub(networks +S*S,. These S*S consist of a series separate -CCs and in(station data communication links, to form a dynamic part of the whole T*7. The important features of S*7 include possessing intelligent network elements and using the embedded -CCs. The combination of the two greatly shortens the T*7 information transfer and response time. *oreover, the network management functions can also be downloaded via -CC to network elements, so as to reali!e the distributed management. It can be said that the basic characteristics of SDH is that it has powerful and effective network management capacity. #ig. /(2= shows the relation between T*7, S*7 and S*S. SES*(7*S can be either an SDH *anagement Subsystem +S*S, or an SDH *anagement 7etwork +S*7,. The relation between it and a T*7 is as follows% &s shown in the above diagram, T*7 is the most general concept of management network. S*7 is its sub(set, specially responsible for management of SDH 7-. S*7 further consists of multiple S*S. &s SES*(7*S is part of T*7, it shall provide standard interfaces to obtain the management from the higher layer network management center. In the SDH system, the logical channel that transfers network management message channels is -CC, and its physical channel should be DCC. It is the /C2kb0s and =B>kb0s channel consisting of the D/(((D3 bytes in the SDH regenerating +regenerate, segment overhead 6SDH and the D;((( D/2 bytes in the multiple$ing segment overhead *SDH, respectively known as DCC+6, and DCC +*,. The former can access trunk stations and terminal stations, while the latter is the

highway for the transmission of network management information between terminal stations.
T*7 S*7 S*S(/ S*S(2 S*S(n

#ig. /(2= Diagram of the relations between S*S, S*7 and T*7

2. SDH management interface The main operation Q running interfaces related to SDH network management are 5 interfaces +containing 53 or 5$ interface, and # interfaces. S*S will communicate with T*7 via the 5 interface.

1.11.3 2M8 +&M8, s"stem function

IT8(T specifies five functions of an 7*S% Configuration *anagement, #ault *anagement, .erformance *anagement, Security *anagement and &ccounting *anagement. Configuration management% configuration of resource and services of a transmission network, including the configuration of network data, e"uipment data, link channel, protection switching function, synchronous clock source allocation strategy, order wire e"uipment, line interface parameter, tributary interface, 7- time, "uery of, backup of and restoration of configuration information, "uery and measurement of path resource, etc. #ault management% detection, analysis and locating of any e"uipment fault, including the setting of alarm levels, real(time display of alarms, the settings of alarm confirmation, shielding, filtering, reversal and sound, "uery of current history alarm, alarm locating, alarm statistics analysis, etc. .erformance *anagement% making an effective detection and analysis of various e"uipment performances, including the setting of performance threshold, "uery of current and history performance data, performance data analysis, etc. Safety *anagement% providing security control for e"uipment maintenance, including the settings of user levels, operation authority and management area, management of user login and that of user operation log, etc. Charging management% providing basic information related to charging, including the set(up time of, duration of, "uality of service of a circuit and so on. Sometimes maintenance management is considered as a separate function module. *aintenance *anagement% providing the means for normal running and fault locating of the e"uipment, including loopback control, alarm insert, error code insert, etc.