Chapter One

Introduction and Basic Concepts

1-1 Introduction.
It has been widely accepted that the 21st century will be an information age, and people will enter the information society. Online letter, online news and surfing that are now commonly used will be extended to online shopping, E-commerce, tele-education, telecommuting, online entertainment, online cinema, tele-medicine, online government, etc. eople!s ways of production and ways of life will then be greatly changed by the highly developed information technology. "omputers and optical fiber communication systems, that use lasers , the acronym of light amplification by stimulated emission of radiation, as source, are the most important technological pillar of the information society. #aser communications present the most exciting, and probably the most challenging, aspects of modern systems. #aser communications found applications in computers, which fall into two ma$or categories, peripheral interconnection and #%& setup. In both areas, one of the critical fiber parameters includes high bandwidth, immunity to electromagnetic interference 'E(I) and electromagnetic pulse 'E( ), and long distance.

1-2 History of Optical Communications.

*he idea of having light guided through bent glass is not new or high tech. #eonardo +a,inci actually mentioned such a means for guiding light in one of his noteboo- .1/. 0owever, he has not been able to verify this assertion. 1hat is -nown for certain is that total internal reflection of light in a beam of water-essentially guided light-was demonstrated by the physicist 2ohn *yndall '1324-1356) in either 1378 or 1394 - depending upon which reference you consult. *yndall showed that light could be bent around a corner while it traveled through a $et of pouring water .1/. :sing light for communications came after this. %lexander ;raham <ell '1389-1522) invented the photo-phone around 1334. <ell

Chapter One

Introduction and Basic Concepts

demonstrated that a membrane in response to sound could modulate an optical signal, light. <ut, this was a free space transmission .1/ system. *he light was not guided. ;uided optical communications had to wait for the 24th century. *he first patent on guided optical communications over glass was obtained by %* =* in 1568. 0owever, at that time there were really no materials to fabricate a glass 'or other type of transparent material) fiber optic cable with sufficiently low attenuation to ma-e guided optical communications practical. *his had to wait for about thirty years. .1/ 1ith the successful development of lasers in 15>4, fiber optics became the technology of future for telecommunications, but problems relating to the two most important elements-the light source and transmission medium-meant the practical applications remained elusive. <y 1594, however, a laser diode capable of continuously lasing at room temperature and an optical fiber with a transmission loss of 24db?-m at last provided the level of performance re@uired. In a little over year, fiber-optic telecommunications underwent remar-able growth .2/. *he first optical cables were put into operation for telephone services in military applications on ships in 1596in the :.A .6/. On Aeptember 1,1599 under the pro$ect named <erlin 1 the first fiber-optic local intercity trun- cable in the telephone networ- was hauled in .8/. #ater, in 1536 and 1538 respectively in the pro$ect named <erlin II, and broadband integrated glass fiber optic local communication networ- '<I;BO&), for the first time optical fibers were run in the homes of the subscribers with the use of graded index fiber .8/. It was in 1535, when high-temperature superconductivity was the rage, that the erbium dopped fiber optical amplifier 'E+B%), an amplifier using E+B pumped by a 1834nm-laser diode, forever changed this perception .2/. In no time the E+B% laser was in actual use, and has

Chapter One

Introduction and Basic Concepts

became the -ey to the technology of wavelength-division multiplexing '1+() .2/. 1+( is the simultaneous transmission of several optical signals of different wavelengths on the same fiber optic cable. It is a techni@ue used so that several different communications channels can share the same fiber optic cable. In 1554 #inn (ollenauer transmitted a 2.7 ;bps signal over a distance of 9.7-m without regeneration using soliton 'solitons are pulses that propagate without changing their shape in optical media .7/) laser and erbium doped amplifier .6/. +uring the early 1554Cs several transpacific and trans-medirerrianean fiber-optic cables were in the process of being installed between the :.A. and Europe, 2apan, and %ustralia, they were routed via fiber-optic cable between continents .>/. *he early 1558 were conducting several field test of fiber to the home conducted in "alifornia and Blorida .9/. %nother success was scored with the optical fiber amplifier 'OB%) repeater systems, for 1+(-based high definition television '0+*,) programming for the winter Olympic in &agano .2/. In 1553,#ucent!s (icroelectronics ;roup created the fiber-optic industry!s first 14-channel tunable transmitter for both short- and long-haul high-speed optical networ-ing systems, using electroabsorptive modulated laser 'E(#) technology pioneered at <ell labs .3/. %t the same date, scientists at <ell labs have made a 24> wavelength laser, a single laser that can do the wor- of 24> lasers by generating light pulses, each lasting $ust 144 millionths of a billionth of a second and using wavelength-division multiplexing 1+( to transmit the data over multiple wavelengths .5/.

1-3

Advantages

and

Disadvantages

of

Optical

Fi er

Communication !ystems.
"ommunication using an optical carrier wave guided along fiber has a number of extremely attractive features, several of which were apparent

Chapter One

Introduction and Basic Concepts

when the optical communication was originally conceived. *he main advantages in using optical fiber communication areD E Enormous potential bandwidth. *his is mainly due to the optical carrier fre@uency, which is in the range 1416 F141> 0G, which yields a far greater potential transmission bandwidth than metallic systems 'coaxial cable bandwidth up to around 744(0G).14/. *he large bandwidth of fiber which indicates the information capacity of a fiber optic cable .1/ ma-es it suitable for bac-bone networ-s .14,11/. Aince the optical fiber communication systems use the optical fiber as a transmission media they have the following advantagesD E #ow transmission loss. Bibers have been fabricated with losses as low as 4.2d<?-m and this feature become a ma$or advantage of optical fiber communication. It facilitates the implementation of communication lin-s with extremely wide repeater spacing 'long transmission distances without intermediate and cost .14/. E Immunity from Electromagnetic Interference. E Electrical Isolation. *he main disadvantages of the optical fiber are.12/D E Ausceptibility of fiber to the evolution of hydrogen and to the ingress of water into the cable. E Ausceptibility to the ioniGing radiation. E #aser and #E+ non-linearities limit analog use. E 0igh @uantum receiver noise. E Bibers are not directly suited to multiaccess use. electronics), thus reducing both system complexity

1-" Aim of t#e $or%.

*his wor- aims to use the laser diode with the optical fiber in highspeed long distance networ-s. In addition it aims to reduce the effects of

Chapter One

Introduction and Basic Concepts

E(I and noise on the performance of the system. *he security of he system is to be increased by this. +ue to the absence of a high-speed serial bus on the computers, the wor- is limited to increase the separation distance between successive stations in the networ-. %lso to cancel the effect of E(I and noise on the system especially in a noisy industrial environment and increasing security of transmission. *his is to be achieved by the utiliGation of the optical fiber as a transmission media and a laser diode as an optical source. *he communication is via the HA-262 standard interface. *he low attenuation factor of the optical fiber as mentioned in the previous section in con$unction with the very intense optical output power ma-e this possible.

1-& Optical Components of Optical Communication !ystem.

Optical communication systems are systems, which use the lasers as carriers for transporting information from the source to the destination. *he optical fre@uencies are part of the electromagnetic spectrum, as shown in the figure '1-1) .16,18/, whose are high fre@uencies in the range of 14 16 to 141> 0G 'i.e. visible light and near infrared waves .17/ with 4.2 to 2 Im wavelength range. *he current theories on electromagnetic radiation state that it has particles 'photons) nature. *hat is photons travelling in a waveli-e fashion. *hat means light has dual nature and its behavior manifasten itself in one of these channels of according to its nature of interaction. *he optical components of a fiber optic communications system as shown in the figure '1-2) .1>,19/ are, in simple terms, a light emitter li-e laser, which initiates the optical signal, a fiber, which transmits it, and a detector, which receives it and converts it into an electrical e@uivalent.

Chapter One

Introduction and Basic Concepts

Figure (1-1) Electromagnetic spectrum.

Connect
Aignal In

+river

#aser

or

+etector

%mplifier and signal

*iming Hecover y

Aignal Out

Biber

Figure (1-2) Basic fiber optic transmission.

1-&-1 'ig#t sources.

#ight sources are a -ey element in any optical fiber system. *hese components convert the electrical signal into a corresponding light signal that can be in$ected into the fiber .14/. *he common light sources for optical fiber systems are the semiconductor light sources, which are light emitting diode '#E+) and laser diode '#+) .13/.

Chapter One

Introduction and Basic Concepts

#aser diodes come in two principal formsD Babry- erot and distributed feedbac- .15/. Babry perot lasers can be further divided according to the lateral confinement of the laser action into gain guided and index guided lasers .12/. ;ain guided device has higher threshold current than index guided one and it has a multimode spectra .24/. (any modes are excited, especially at power levels not far above the threshold. *hus, a rather broadband spectral width is obtained. 1ith index guided lasers several modes are usually present $ust above lasing threshold as the driving current increased, one or two modes cause to dominate the others .12/. % comparison of the advantages of #E+ and #+ is shown in table '1-1) .15/. #ight-emitting diodes may be used with either multimode or single-mode fibers when lower light levels and a lower Information capacity 'lower modulation bandwidth) are acceptable. #aser diodes achieve higher performance at the expense of higher cost and complexity .21/. Table (1-1) Comparison of light sources.
+esirable "haracteristics #ight-Emitting diode high launched optical power Aimple electronics low cost &arrow linewidth #ow Aimple #ow 1ide Babry- erot #aser 0igh less simple Intermediate &arrow 0igh +istributed-Beedbaclaser high less simple high very narrow high

0igh modulation bandwidth #ow

1-&-1-1 'aser Diode Operation

#aser diode is a solid state-devise that emits a coherent, very intense beam of light over a narrow bandwidth .22/. *his results in a higher data rate over longer distances. Its & $unction forms an optical cavity, which is made to laser .26/. #aser diode is similar to the #E+ below a certain

Chapter One

Introduction and Basic Concepts

threshold current levels which is defined as the current at which the laser action is started. %bove the threshold, lasing action is started .22/ that is when the in$ected current is so large that population inversion ta-es place .22/. *he $unction is forward biased and emits photons. Aome photons JhitJ free electrons that then recombine and emit coherent photons by simulated emission. *hat is when an incident photon leads an electron to lose an amount of energy, then a second photon is emitted with the same fre@uency and phase as that of the first photon. *hese simulated emissions form a chain reaction. *he chain reaction is self-sustaining if the gain due to simulated emission is larger than the loss due to the imperfect reflection and to absorption. .26/. #aser diodes re@uire cooling and consume more power than do #E+s. 0owever, they offer much more bandwidth, as they can pulse on and off in a fraction of a nanosecond. Burther, contemporary diode lasers create light fre@uencies, which are attenuated much lessK hence, the signal can travel much farther without being repeated. *here are several different schemes for carrying out the modulation function. *hese are respectively .26/D 1. Intensity modulation 'I() which modulates the amplitude of the optical carrier. 2. Bre@uency Ahift -eying 'BAL) which fre@uency modulates the optical carrier. 6. hase Ahift Leying ' AL) which phase modulates the optical carrier. 8. olariGation (odulation ' () which modulates the olariGation Atate of the optical carrier. 1ithin the context of a fiber optic data lin- the only one really employed is Intensity (odulation I( .1/. (ore details about this techni@ue is introduced in appendix '%).

Chapter One

Introduction and Basic Concepts

1hen considering the use of the diode lasers for optical communications, it is necessary to be aware of certain of its characteristics, which may affect its efficient operation .14/. Aome of the most important characteristics of the laser diode are illustrated in bellow.

1-&-1-2 (#res#old current temperature dependence.

*he temperature dependence of the threshold current density is approximately exponential for most common structure .1>/. *he increase in threshold current is due to the increasing spread of electrons and holes In$ected into the conduction and valance bands .14/.

1-&-1-3 !pectral )roperties.

One important property of a #+ is the wavelength spread over which the optical power is distributed. *he width of the spectrum emitted by a light source is the bandwidth of the source .1/. *he spectral width is the 6-d< optical power width 'measured in nm or microns). *he spectral width impacts the effective transmitted signal bandwidth. % larger spectral width ta-es up a larger portion of the fiber optic cable lin- bandwidth. Bigure '1-6) illustrates the spectral width of the #+ versus the #E+ .1/. *he spectral width is the half-power spread. *he specific value of the spectral width depends on the details of the diode structure and the semiconductor material. 0owever, typical values for an #E+ are around 84 nm for operation at 374 nm. *ypical values for a #+ are 1 nm for operation at 374 nm .1/. %nother important property of the laser diode is the dependence of its wavelength on the temperature 'about 4.27 nm?-M) .1>/ and on the in$ection current 'about 4.47 nm ? m%) .14/.

Chapter One

Introduction and Basic Concepts

10

Figure *1-3+, '-D and laser diode spectral .idt#

1-&-1-" Optical )o.er.

*he radiant output power versus the in$ected current of a typical laser diode is shown in figure '1-8) .28/. It can be seen that a very little output power is realiGed until the threshold is reached, at which the lasing occurs. %bove the threshold current point, a great change in the output optical power with a small change in in$ection current is obtained.

Chapter One

Introduction and Basic Concepts

11

7 peak laser power output 6 5 4 3 2 1 0 0 100 200 300 peak current (mA)

Figure (1-4) The radiant output po er !ersus the in"ected Current of a t#pical laser diode.

1-&-2 Optical Fi er.

*he optical fiber is a thin '2 -127 (m), flexible medium capable of conducting optical signals. ,arious glass and plastics can be used to ma-e optical fiber .16/. %n optical fiber consists of a central section, made of silica .18/, called the core which is the light-transmitting region of the fiber and a cladding made of pure silica, the first optical layer around the core .18/. ropagation of light in fibers can be dealt with in term of geometric optics or wave optics. In geometric optics propagation of light can manipulate in tem of ray optics and total internal reflection. In wave optics it interpreted in term of modes. *hese modes are those electromagnetic waves that satisfy the homogeneous wave e@uation in the fiber and the boundary condition at the waveguide surface. *o analyGe the optical waveguide, one needs to consider (axwellCs e@uations that give the relationships between the electric and magnetic fields. %ssuming a linear, isotropic dielectric material having no currents

Chapter One

Introduction and Basic Concepts

12

and free charges. % simple analysis of optical waveguide in a cylindrical waveguide is given in appendix '<) .17/. *here are two basic types of optical fiber, characteriGed by the paths 'modes) through which the optical rays guided as they travel through the fiber, they are single-mode and multi-mode, Aingle-mode fiber allows only one mode of light to propagate through the core. (ulti-mode fiber allows hundreds of rays of light to propagate through the fiber simultaneously. ropagation of light along a waveguide can be described in terms of set of guided modes of the waveguide. Each guided mode is a pattern of electric and magnetic field lines that is repeated along the fiber at intervals e@ual to the wavelength. Only a certain discrete number of modes are capable of propagating along the guide. Brom the point of view of system designer, the most important characteristics of an optical fiber are the fiber losses and the dispersion. *hese two factors cause a degradation of the light signal that is launched into the fiber as shown in figure '1-7).1>/. *he informationcarrying capacity of a glass fiber is limited by the degree of these factors .27/.

Figure (1-$) Causes of signal degradation in optical fiber.

Chapter One

Introduction and Basic Concepts

13

1-&-2-1 Fi er 'osses.
Biber losses set the maximum fiber length in term of the available light power .18/. *he losses in the optical fiber are dominated by the impurities in the silica-based materials that are used for fabricating fibers li-e hydroxyl impurity O0 .16,2>/. *he loss of light power that occurs as light propagates through the fiber is expressed in decibels per -ilometer d<?-m as in e@uation '1-1) .16/. ower loss N -14 log ' owerout? ower in)OOOOO..'1-1) *he fiber loss varies dramatically with wavelength. %typical example is shown in the figure '1->) .13/. *his figure shows the schem of loss as a function of wavelength P for pure silicon fiber. *he attenuation curve shows that minima at wavelengths of 374 nm, 1644 nm, and 1774 nm they form the three operating windows. *he wavelength at which minimum attenuation occurs is close to 1774 nm. 0ence, the amount of attenuation caused by an optical fiber is primarily determined by its length and wavelength of the light travelling through the fiber .6/.

Chapter One

Introduction and Basic Concepts

14

Figure (1-%) Fiber loss !ersus

a!elength.

1-&-2-2 Dispersion.
+ispersion is the technical term for the spreading of the pulses of light as they travel down the fiber. .6/. +ispersion can be divided into material dispersion and chromatic dispersion. (aterial dispersion refers to variation in the group velocity of light in glass with wavelength .29/. "hromatic dispersion represents the fact that different colors or wavelengths propagate at different speed .29/. *he dispersion sets the maximum fiber length or repeater spacing as a function of signal fre@uency or data bit rate .18/. +ispersion limits bandwidth in optical fiber, reducing the amount of the information that the optical fiber can carry .6/. *able '1.2) shows simple comparison of multimode and single-mode optical fibers .15/. Table (1-2) Comparison of multimode and singlemode optical fibers.
Desira le C#aracteristics high information capacity over distance /ultimode Fi ers !ingle-/ode Fi ers low 0igh

low cost Aimple interconnections

intermediate simple

#ow +emanding

1-&-3 Optical Detector

% detector is an opto-electric device that converts the received optical power into electrical power with linear response .14/. *here are two principal types of detectors used in optical-fiber communications systems. <oth are Aemiconductor devices, they are positive-intrinsic-negative photodiodes ' I&s) and avalanche hotodiodes '% +s) .23/. % comparison between these types is shown in table '1-6) against a list of desirable

Chapter One

Introduction and Basic Concepts

15

characteristics for detectors .15/. ;enerally spea-ing, the I& photodiodes have lower cost and lower noise. 0owever, they offer no gain. *he avalanche photodiodes have high gain but also higher noise. Table (1-&) Comparison of detectors.
+esirable "haracteristics low cost Aimple electronics low temperature sensitivity high light sensitivity low internal noise I& low simple low low low % + high complex high high high

*he output of the detector is a very wea- signal, so the photodiode circuitry is followed by one or more amplification stages. *his is to amplify the signal to yield an electrical signal representing the original input .25/. *here may be even filters and e@ualiGers to shape and improve the information bearing electrical signal .1/. *he complete Heceiver must have high detectability, high bandwidth and low noise. It must have low noise so that it does not significantly impact the bit error rate '<EH) of the lin-. *his is the probability that a transmitted bit is demodulated in error at the destination receiver .1/. (ore details about direct detection are introduced in appendix '%).

1-0 'in% 1udget.

+esigning of an optical fiber lin- is not trivial, however, there are logical ways to proceed. One way is to first analyGe the optical lin- power budget. *his is to ensure that sufficient power is available throughout the lin- to meet the demands of a given application. In addition, ade@uate bandwidth for a system can be assured by developing a rise time budget.

Chapter One

Introduction and Basic Concepts

16

Optical )o.er 'in% 1udget. In order to ensure satisfactory performance of the optical communication system, the designer must budget the optical loss in the system. *he power budgeting for a digital optical fiber communication system is performed in a similar way to power budgeting within any communication system. *he power budget is the calculation of the power received by the photodetector. *he calculation subtracts the power loss during transmission expressed in decibel 'd<) from the transmitted optical power expressed by d<m 'decibel with relative to miliwatt). *he difference is the received power in d<m .22/. It is necessary to incorporate a system margin into the optical power budget so that small variations in the system operating parameter do not yield an unacceptable decrease in system performance. *he safety margin '(a) depends to a large extent on the system components as well as on the system design procedures. *his margin is typically in the range 7-14 d< .14/. #etting the average optical power re@uired by the photdetector to maintain a proper A&H or <EH is
r

, the optical power launched into the

fiber emitted is ps, and the channel losses be #. *hen, the optical power budget for a system is given by the following expression .14/D N Q # Q(a OOOOOOOOO'1-2)

*he channel losses can be written as .14/ # N 'R Q R$) + Q Rc OOOOOOO...'1-6) whereD R is the fiber attenuation in decibel per -ilometer d<?-m,R$ is the losses due to the $oints 'generally splices) in decibel per -ilometer d<?-m, Rc is the connector loss in d< which are used for coupling the optical source and detector to the fiber, and + is the power limited

Chapter One

Introduction and Basic Concepts

17

distance of the lin-. <eyond the distance +, the received power is not sufficient to maintain the re@uired A&H or <EH. 2ise (ime 1udget. Effective transition time ts of the optical signal entering the receiver should be less than 94 percentage of the signal bit rate <. *hat is ts N 4.9<OOOOOOOOOOOOOOOO.'1-8) *he system rise time is obtained as in e@uation 1-7. ts N 'tl2 Q tm2 Q tc2)1?2 OOOOOOOOOOOO'1-7) whereD te2 is due to transmitter electronics, tc2 is due to diode laser, and tm2 is the time transition due to modal dispersion. *ransition time, which refers to both rise and fall times, is e@ual to '4.88# @ ? <f), whereD @ is a constant factor that indicates hoe modal dispersion scales with length, < f is the fiber bandwidth in (0G?-m. @ is 4.9 for long fiber distances and 1.4 for short fiber distances. tc is the chromatic dispersion and usually neglected. %s long as ts is shorter than this value, the full system bandwidth is limited by the receiver. If it is greater, the performance of the lin- is limited by an element other than the receiver. *hat is the sensitivity of receiver will decrease for a given <EH value.

1-3 Common (ransmission !tandards.

*here are many types of fiber optic data lin-s. *he variety of lin-s reflects a wide range of applications, data rates, and formats. Bigure 1-14 shows many of the transmission standards now in use .6/.

Chapter One

Introduction and Basic Concepts

18

+ata rates from HA-262 to HA-822?737 are short-haul data rates while European "E *?""I**, 2apan &I*, ""I** <E##?""I**, +A1-+A8 and AO&E*, O"1-O"83 relate to current telecommunication lin-s.

Figure (1-') (orld ide communication standards *he transmission standard EA"O& is gaining popularity as a means to connect computers to peripherals, and computer to computer. %ll transmission standards are arranged from lowest 144b?s to highest (ost available fiber optic data lin-s are designed to handle a single data protocol. Hecently, fiber optic data lin-s, capable of meeting the e@uipments of several data protocols simultaneously, have been begun to appear. *hese are -nown as multi-protocol lin-s .6/. *he B++I is the networ-ing technology that this wor- will deal with.

1-4 5et.or%ing (ec#ni6ues.

*he main techni@ues in ma-ing a networ- are .25/D

Chapter One

Introduction and Basic Concepts

19

Ethernet. It is the most widely used networ-ing technology where computers either are connected to a networ- or a common hub. Biber +istributed +ata Interface 'B++I). It is a ring-based networ-, where computers are connected either to a networ- or a networ- hub or onto a fiber ring. IA+&. *his is used to connect computers over a telephone connection.

1-7 Fi er Distri uted Data Interface.

Biber +istributed +ata Interface is a standard for networ- techni@ue based originally on fiber optics. It was established by %&AI '%merican &ational Atandards Institute) in the mid 1534s. *oday, its specifications also include standard copper media for interconnection. B++I basically represents the technical specifications for a #ocal %rea &etwor- '#%&) to communicate at a standard rate of 144(bs but today is also widely used to interconnect #%&s and 1ide %rea &etwor-s '1%&s) over longer cable lengths .25/. <ecause of its superior speed, B++I is especially useful for sending large files such as graphics or digital video. It is a useful method of adding bandwidth without having to ma-e a costly, long-term commitment .61/.

1-7-1 FDDI C#aracteristics and !pecifications.

B++I supplies higher degree of reliability and security. *his is due to the use of the optical fiber as a transmission medium and the physical topology of the B++I. *he advantages of using optical fiber are illustrated in section '1-6). B++I specifies the use of dual separate rings, the ring connects computers on a single circle of cable, around which data moves simultaneously in opposite directions .25, 62,66,68/, as shown in figure '1-3).67/. One ring is the primary and the other is the secondary or

Chapter One

Introduction and Basic Concepts

20

bac-up data ring .6>/. *he secondary ring is redundant and is used for fault recovery .25, 69/. It used only in the case of failure of the primary ring or attached wor-station .66/. B++I uses the timed to-en as an access method .67/. % to-en is a special series of bits that travels around a to-en-ring networ-. Each networ- has only one to-en .67/.

Figure 1.) F**+ uses a dual-ring topolog# %nother important characteristic of B++I is its ability to interoperate easily with IEEE 342.6 14 (bps Ethernet networ-s. It is used as a bac-bone to large networ-s, as well as an interim networ- for connecting #%&s to high-speed computers. %n B++I-to-Ethernet bridge is the specific technology employed in such a map. <ridges may be able to connect either a single Ethernet or several Ethernet to the B++I networ- .66/. *he B++I capacity allocation scheme see-s to accommodate a mixture of stream and bursty traffic. B++I defines two types of trafficD synchronous and asynchronous. Each station is allocated a part of the total capacity 'the portion may be Gero)K the frames that it transmits during this time are referred to as synchronous frames. %ny capacity that is not used is available for the transmission of additional frames referred to as a synchronous frames .63/.

Chapter One

Introduction and Basic Concepts

21

One of the uni@ue characteristics of B++I is that multiple ways actually exist by which to connect B++I devices as in figure 1-12. B++I defines the following types of devices .64,67,63/D Dual Attac#ed !tations *DA!+ such as servers or routers that are attached to both primary and secondary rings and are involved in ring wrap when ring faults occur. !ingle Attac#ed !tations *!A!+. It is attached to ports of a concentrator and only on the primary ring. Dual Attac#ed Concentrators *DAC+. *hese provide a reliable connection for stations to the bac-bone. !ingle Attac#ed Concentrators *!AC+. *hese are less reliable due to the single connection to the bac-bone. &ull %ttached "oncentrators '&%") are not connected to any bac-bone. *he basic B++I specifications are listed in table '1- 8).25/. Aome of B++ICs benefits are .6>/. 1. 0igher "apacity and erformance than older #%&s. 2. (ore Aimultaneous *ransactions. 6. 0igher %vailability 'dual ring topology). 8. redetermined erformance 'adding users have minimal impact on throughput). 7. #onger +istance loops '2 -ilometers to 144 -ilometer).

Chapter One

Introduction and Basic Concepts

22

Figure (1-,) -arious

a#s to connect F**+s

de!ices. (a le *1- "+ t#e asic FDDI specification

)A2A/-(-2 *opology %ccess method *ransport (edia (ax Brame AiGe *ypical Hate (aximum &umber of Atations (aximum +istance <etween stations %ttenuation <udget D-!C2I)(IO5 Hing *imed *o-en Optical fiber, shield twisted pair, unshield twisted pair 8.7Lb 144(bps 744 each ring'1444) 2-m'for multimode fiber cable4,>4-m'for single-mode fiber) 11db'between stations), 1.7db?-m at 1644nm for >2.7?127 fiber

1-18 !tation Organi9ation.

*he B++I covers the first two layers of the open system interconnection 'OAI) model. B++I is not actually a single specification, but it is a collection of four separate specifications each with a specific function. "ombined, these specifications have the capability to provide high-speed connectivity between upper-layer protocols such as transport

Chapter One

Introduction and Basic Concepts

23

control protocol internet protocol '*" ?I ) and internet pac-et exchange 'I S), and media such as fiber-optic cabling. Bigure '1-14) illustrates the four B++I specifications and their relationship to each other and to the IEEE-defined #ogical-#in- "ontrol '##") sublayer .29,68/. *he ##" sublayer is a component of #ayer 2, the data lin- layer, of the OAI reference model .64,61,69,65/. *hese specifications areK hysical #ayer (edium +ependent ' (+) hysical #ayer rotocol ' 0T) (edia %ccess "ontrol '(%") Atation (anagement 'A(*) #ogical #in- "ontrol Aublayer '##").
OAI B++I #ogical #in- #aywer

+ata #in#ayer

(edia %ccess "ontrol

Atation (anagement

hysical #ayer

hysical #ayer rotocol

hysical #ayer (edium

Figure (1-1.)/ F**+ specifications map to the 01+ hierarchical model

Chapter One

Introduction and Basic Concepts

24

1-18-1 )#ysical 'ayer /edium Dependent *)/D+. *his defines

the characteristics of the transmission medium, including the fiber-optic lin-, power levels, bit error rates, optical components, and connectors .25,66/. *he standard is IAO 5618-6 (+ .67/. (+ defines "able types and associated construction and connectors %ttenuation budget Optical bypass engineering considerations *ransmitter and receiver specifications 1644 nm wavelength (ultimode fiber are with >2.7?127 Im core cladding diameters and 4.297 numerical aperture. Biber has a <EH of 14-12. *he Overall networ- <EH of 14 -5 .6>/. *he transmitters are #E+s and the receivers are I&-diodes .23/. Aingle-mode fiber. Biber optic cable with a core siGe of 9 to 11 Im .6>/. It allows distances up to 14-m .64/. *he transmitters are lasers .14/. *he A(B- +( will enable individual lin-s up to >4-m or even up to 144-m.14/.

In bellow the specification of the +( different media

1-18-2 )#ysical 'ayer )rotocol *)H:+.

0T encodes and decodes symbols that are the smallest pieces of information between the (%" and the (+ using the 8<?7<-encoding scheme. In this scheme, each group of 8 data bits encoded in a 7 bits code group called a symbol. *his encoding allows 34 percentage efficiency since it ta-es 7 bits to convey 8 bits of information. Aymbols are defined for the hexadecimal numbers 4 through B and also for a number of communication-specific functions. *able '1-7) shows the symbols used and their 8<?7< encoding. *he 7 bit symbols are selected so as to guarantee a line transition at least every 6-bit time on the medium .69/. *he actual data transmission of the data is accomplished using the nonreturn to Gero

Chapter One

Introduction and Basic Concepts

25

inversion '&HUI) transmission scheme rather than normal binary transmission. *hat is a signal transition occurs when a 1 is transmitted and no transition when a 4 is transmitted. *he use of 8<?7< coding and &HUI ensures a signal transition every 2 bits .66,68/. *he physical interface is depicted in figure '1-11).61/. *he standard is IAO 5618-1 0T .6>/.

Brom (%" :nit

*o (%" :nit

8<?7< Encoder

127 (0G #ocal "loc-

8<?7< +ecoder

&HUI Encoder

"loc- Aynchronisor

&HUI +ecoder

Optical *ransmitter

Optical Heceiver

*ransmission "ontrol :nit

BiIber Out

BiIber In OuWIn t

Figure (1-11) schematic of F**+ ph#sical interface. Table (1-$) F**+ 4B2$B data s#mbols and control s#mbols.
0exadecima l 4 1 2 6 8 7 > 9 8bit data 4444 4441 4414 4411 4144 4141 4114 4111 7bit data 11114 41441 14144 14141 41414 41411 41114 41111 hexadecima l 3 5 % < " + E f 8bit data 1444 1441 1414 1411 1144 1141 1114 1111 7bit data 41114 41111 14414 14111 11414 11411 11144 11141 "ontrol 7-bit symbol I+#E 2 L * H A V:IE* 0%#* symbol 11111 11444 14441 41141 44111 11441 44444 44144

1-18;3 /edia Access Control */AC+.

Chapter One

Introduction and Basic Concepts

26

*he functions of (%" sublayer are manage data transfers, error recovery, networ- addressing, to-en passing, and station access. *he standard is IAO 5618-2 (%" .6>/.

1-18-3-1 /AC Frame Format.

*he contents of this format are defined in terms of symbols, with each data corresponding to 8 data bits. Bigure '1.17a) depicts the frame format for he proposed protocol .63,65/. % frame other than to-en frame consists of the following fieldsD )ream le *)A+. AynchroniGes the frame with each stationCs cloc-. .65/. It is comprised of 1> or more I+#E 'nondata) symbols which, since they each consists of five 1bits. *hey cause the line signal to change at the maximum fre@uency .63/. !tarting delimiter *!D+. *he start delimiter field consists of two control symbols '2 = L) which enable the receiver to interpret following frameCs contents on the correct symbol boundaries. .63/ Frame Control *FC+. *he frame control field defines the type of the frame '(%" or information) and certain control functions .63/. It has the bit format "#BBUUUU, whereD C indicates whether this is asynchronous or synchronous frame. "N1 BOH synchronous frame and "N4 for asynchronous frame. ' indicates the use of 1>bit or 83bit address. FF Indicate whether this is an ##", (%" control, or reserved frame. Bor a control frame, the remaining 8bits indicate the type of control frame. .69/ Destination Address *DA+. It specifies the station's) for which the frame is intended. It may be a uni@ue physical address, a multicast-group address or a broadcast address .65/. It can be either 1>bits or 83bits long, but for any specific #%& they are the same for all data terminal e@uipments

Chapter One

Introduction and Basic Concepts

27

'+*Es). *he first bit of the field indicates whether the address is an individual address '4) or a group address '1). Individual address identifies a specific +*E on the ring while group addresses are used to send a frame to multidestination +*Es. In addition, a +% consists of all 1s is a broadcast address denoting that the frame is intended for all +*Es on the ring .63/. !ource Address *!A+. *his specifies the station that sent the frame .b68p821/. *he A% can be either 1>bits or 83bits long. *he A% is always an individual address. .63/ Information. *his field contains an ##" data unit or information related to a control operation .65/. %lthough no maximum length is specified for the information field, it is limited in practice by the maximum time for which a +*E is allowed to transmit a frame when holding the control to-en. .6>/ % typical maximum length is 8744 octets .69/, including data and all frames "omponents .65/. Frame C#ec% !e6uence *FC!+. *he frame chec- se@uence is a 62 bits cyclic redundancy chec- '"H"), based on the B", +%, A%, and information fields .63/. B"A is filled by the source station with a calculated value dependent on the mentioned frame contents. *he destination station recalculates the value to determine whether the frame was damaged in transit. If so, the frame is discarded. -nding Delimiter *-D+. It contains a nondata symbol '*) and mar-s the end of the frame, except for the frame status field .65/. Frame !tatus *F!+. "ontains the error-detected 'E), address-recogniGed '%), and frame copied '") indicators. Each indicator is represented by a symbol, which is H for XresetY or XfalseY and A for XsetY or XtrueY.65/. % to-en frame consists as shown in figure 1-17b of the following fields .65/D !tarting Delimiter *!D+. Is %s illustrated above.

Chapter One

Introduction and Basic Concepts

28

Frame Control *FC+. 0as the bit format 14444444 or 11444444 to indicate that this is a to-en. -nding Delimiter *-D+. Is %s illustrated above.

a, FDDI frame.

b: token frame. Figure (1-1$) F**+ frame format of a/ F**+ frame. b/ To3en frame. 1-18-3-2 (imed (o%en )rotocol *he proposed protocol uses timed to-en mechanism where a station can hold the to-en no longer than a specified amount of time. *herefore, there is a limit to the amount of data that a station can transmit on any given opportunity .25/. It is similar to that used with a to-en bus networ.25, 63/. *he protocol is controlled by a preset parameter -nows as the target to-en rotation time '**H*) .63/. *he time that the station holds a to-en as called the to-en holding timer '*0*). *he rotating time '*H*) defines the time that a to-en ta-es to traverse around the ring. It can be seen that .25/D &odesZ*ransmitting [ *0* Q HingZ#atency \N *H*OO'1->) whereD &odesZ*ransmitting defines the number of nodes that re@uire transmitting data and HingZ#atency is the amount of time it ta-es data to traverse around the ring when there are no stations transmitting data .25/.

Chapter One

Introduction and Basic Concepts

29

%fter initial negotiations, all stations on the ring agree to release the to-en within the **H*. % to-en travels a round the ring and is captured by a particular station until all data has been sent or until the *0* expires. Each station on the ring chec-s each frame it receives for errors and retransmits it. *he destination station copies the frame it receives and sets the error indicator, the frame copied indicator, and the address recogniGed indicator. *he destination station sends the frame on until the sending station receives it bac-. %t this point, the originating station ta-es the frame from the ring once it recogniGes its own address in the source address field. 1hen the station which sent the frame strips it from the ring, it examines the % and " bits. *he three possible combinations areD %N4 and "N4K +estination not present or powered up. %N1 and "N4K +estination present but frame not accepted. %N1 and "N1K +estination present and frame copied.

*his arrangement provides an automatic ac-nowledgment of the delivery status of each frame. *he timed to-en protocol ensures that access to the ring is shared fairly between all stations by ma-ing all stations on the ring -eep a note of the *H*. %s the load on the ring increases, this time will increase. *he *H* is compared with **H*. *he stations are allowed to transmit any waiting frames only if its measured *H* '(*H*) is less than the preset **H* for the ring. 0ence, when a station has frames waiting to send, on receipt of the to-en it computes the difference between (*H* and **H* and transfer it to the *0* which determines for how long the station may transmit waiting frames before releasing a new to-en. If *0* is positive, then the station can transmit for up to this interval. If it is a negative, then the station foregoes transmitting any waiting frames for this rotation of the to-en. % positive (*0* is thus -nown as an early to-en while a negative (*0* is -nown as a late to-en .25/.

Chapter One

Introduction and Basic Concepts

30

% case can occur when a station receives a to-en, and the (*H* is less than the **H*, but the difference in time is not enough to transmit the whole data frame. In this case, the station is allowed to transmit it. *hus the (*H* will actual have a maximum value of the *H* plus the total time to transmit a full frame.

1-18-" !tation /anagement *!/(+.

*his is used to communicate between stations to maintain and monitor the ring by collecting statistics detecting, isolating and solving faults .25,63/. *he A(* protocol handles the management of the B++ICs ring, which include .25/D %d$acent neighbor indication. Bault detection and reconfiguration. Insertion and deletion from the ring. *raffic statistics monitoring. *he ring must be set before normal operation can ta-e place. If a +*E wishes to $oin an already operational ring, that +*E must first go through an initialiGation procedure to ensure that it does not interface with the correct functioning of the established ring. In addition during normal operation it is necessary for each active +*E on the ring to monitor continuously its correct operation. "ollectively these functions are -nown as ring management. *hese functions are illustrated briefly in below .25/D Initiali9ation. (he initialiGation procedure starts with the transmission of a duplicate address test '+%*) (%" frame by the +*E. *he +%* (%" frame is used to ensure that no other +*Es in the ring are using the same address. %fter the +*E ensured that no other station on the ring uses its own address, the +*E transmits a standby monitor present 'A( ) (%" frame.

Chapter One

Introduction and Basic Concepts

31

*he A( (%" frame is used for updating the upstream neighbor address ':&%) in the ring. *he :&% is re@uired for fault detection and monitoring functions. %fter completion of the initialiGation se@uence, the +*E can start to transmit and receive normal frames and to-ens. !tand y /onitor !tate. *he +*E enters the standby monitor state to monitor continuously the correct operation of the ring. It does this by monitoring the passage of to-ens and special active monitor present '%( ) (%" frames, which are periodically transmitted by the currently active monitor. If to-ens or %( frames are not detected periodically, the standby monitor timer times out and the +*E enters the claim to-en state. Claim (o%en !tate. In the claim to-en state, the +*E continuously transmits claim to-en '"*) (%" frame. *he "* (%" frame is used to determine a new active monitor if the current one fails. Active /onitor. If a +*E is successful in its bid to become the new monitor, it initiates the transmission of a purge ' H;) (%" frame to ensure that there are no other to-ens or frames on the ring before it initiates the transmission of a new to-en. %s the ring has been successfully purged, the +*E initiates the neighbor notification process by broadcasting an %( (%" frame. %fter a short delay, this is followed by the transmission of a new control to-en.

1-11 'ogical 'in% Control *''C+ !u layer.

*he logical lin- layer '##") is the upper sublayer of the data linlayer in the OAI system. *he ##" sublayer manages data-lincommunication and defines the use of logical interface points called service

Chapter One

Introduction and Basic Concepts

32

access points 'A% ). Other computers can refer to and use A% s to transfer information from the ##" sublayer to the upper OAI layers .63/. *he ##" sublayer accepts the userCs data unit 'networ- protocol data unit & +:), the +estination A% '+A% ), the source A% 'AA% ), +%, and A% from the upper layer. It arranges the frame and passes a re@uest to the (%" sublayer together with the +%, A%, and logical lin- layer protocol data unit '##" +:). *he frame format of ##" sublayer is shown in figure '1-1>).65/. *he connection between the (%" sublayer and the rest of the communication system is through a structure called a service access point 'A% ). *he A% is a one-byte field. *he service access points, both destination and source as shown in figure '1-1>), indicate which upper-layer protocol the frame is intended for. *he ##" sublayer would loo- at the destination service access point '+A% ) value to determine where to place the frame for further processing. *he protocol specified by the A% s assigned hexadecimal values that some of them are listed in table 1-1>.84/. Bigure '1-19) shows the relationship between the ##" sublayer and the (%" sublayer .81/. +A% AA% control

Figure *1-10+ ''C frame formats.

Chapter One

Introduction and Basic Concepts

33

Figure (1-1') +EEE 44C and 56C la#ers.

1-12 Dra. ac% of FDDI.

*he important drawbac- of B++I is the price. *he B++I networadapter range from ]1744 to ]9744. In turn, these prices with the average Ethernet card at ]97- ]244 and the average to-en ring card at roughly twice that price .62,82/.

1-13 )erformance of t#e FDDI.

*here are two important performance measures associated with shared-medium networ-sD the maximum obtainable throughput and the maximum access delay. <oth are strongly influenced by the (%" algorithm used to share the available transmission capacity between the

Chapter One

Introduction and Basic Concepts

34

attached stations .63/. %lthough the maximum throughput of an B++I ring is 144(bps, because of the ring latency and access control mechanism, the maximum obtainable throughput is less than this. *he maximum utiliGation of the nominal ring capacity :max can be expressed as .63/D :max N **H* F *l ? '**H* F *l) Q *t Q '*l ?n)OOO..'1-9) where *l is the ring latency, *t is the time to transmit the to-en, and n is the number of stations in the ring. In practice, the **H* is very much greater than *t. 0ence the maximum utiliGation can be approximated to .63/D :max N n '**H* F *l) ?'n **H* Q *l)OOOOOOO'1-3) Brom above e@uation, to achieve higher level of ring utiliGation, one must select a **H*, which is significantly greater than the total ring latency. *he access delay is defined as the time delay between the arrival of a frame at the ring interface of the source station and its delivery by the ring interface at the destination station. *he general expression used to compute the maximum access delay of an B++I ring %max is given by .63/D %max N 'n-1)'**H* F*l) Q n*l Q*l N 'n-1) **H* Q2*l OOOOOOOO..OOOOO..'1-5) whereD %max is the worst case time to receive a usable to-en. Aince the ring latency is fixed for particulars ring configuration, the larger the ring **H*, the larger the maximum access delay.

1-1" FDDI Over Copper.

Chapter One

Introduction and Basic Concepts

35

%n alternative to fiber based B++I is to run B++I over copper wiring, either shielded or unshielded twisted pair ':* ), as used in Ethernet and to-en ring installation .84/. "ost saving of :* over fiber amount to about 66^.62/. *his variation of B++I has been dubbed copper distributed data interface '"++I). "++I provides data rates of 144 (bps and uses dual ring architecture to provide redundancy .88/. *he "++I supports a limited distances to 144m?seg as compared with 2-m?seg of fiber based B++I. It is consistent with the physical and (%" layers defined by the %&AI standard .88/.