Local Area Networks

TELECOMMUNICATIONS

Laboratory

LAN & INTERNET

Volume 1/3 Theory

TEACHER / STUDENT handbook

LAN$01E1.DOC

CONTENTS

CONTENTS 1.

2.

INTRODUCTION to LOCAL AREA NETWORKS 1.1 WHAT IS a LOCAL AREA NETWORK 1.2 ELEMENTS of a LOCAL AREA NETWORK 1.2.1 Network technology 1.2.2 Network users 1.2.3 Network servers 1.3 TRANSMISSION MEANS 1.4 NETWORK ARCHITECTURES 1.4.1 Bus topology 1.4.2 Star topology 1.4.3 Tree topology 1.4.4 Braid topology 1.4.5 Ring topology 1.5 QUESTIONS

1 1 3

5 5

9

TRANSMISSION MEANS and CABLING SYSTEMS 2.1 MULTIPAIR cable 2.2 COAXIAL CABLE 2.3 OPTICAL FIBER 2.4 WIRELESS 2.5 CABLING SYSTEMS 2.5.1 Proprietary cabling 2.5.2 Structured cabling 2.6 The STANDARD EIA/TIA 568 2.6.1 The topology 2.6.2 Cabling elements 2.6.3 The transmission means 2.6.4 The backbones 2.6.5 The horizontal cabling 2.6.6 Cables identification 2.6.7 Documents 2.6.8 Types of connectors and splicing systems 2.7 QUESTIONS

10 10 12 13 16 16

3.

NETWORK ACTIVE COMPONENTS 3.1 NETWORK CARD (Network Interface Card - NIC) 3.2 TRANSCEIVER 3.3 HUB 3.4 REPEATER 3.5 BRIDGE 3.6 SWITCH 3.7 ROUTER 3.8 QUESTIONS

28 28 29 31 32 33 34 35 36

4.

SIGNAL CODING TECHNIQUES 4.1 INFORMATION TRANSMISSION 4.2 NRZ (Non Return to Zero) and RZ (Return to Zero) CODING 4.3 NRZI (Non Return to Zero Inverted) CODING 4.4 MANCHESTER AND DIFFERENTIAL MANCHESTER CODING 4.5 MLT-3 41 CODING 4.6 4B5B - 5B6B – 8B6T CODING 4.7 QUESTIONS

37 37 39 39 40

17

26

42 43

CONTENTS 5.

ACCESSES CONTROL TECHNIQUES 5.1 ACCESS METHODS 5.2 Carrier Sense Multiple Access / Collision Detection (CSMA/CD) 5.3 TOKEN-PASSING 5.4 TOKEN-BUS (IEEE 802.4) 5.5 QUESTIONS

45 45 46 48 48 49

6.

INTERNATIONAL STANDARDS 6.1 NETWORK ARCHITECTURE 6.2 ISO/OSI MODEL 6.1.2 OSI frame construction 6.3 LAN IEEE MODEL 6.3.1 Level 2 – LLC sub-level (Logical Link Control) 6.3.2 Level 2 – MAC sub-level (Media Access Control) 6.3.3 Level 1 - Physical 6.4 REFERENCE MODELS 6.4.1 IEEE 802 6.4.2 IEEE 802.3 6.4.3 IEEE 802.5 6.4.4 FDDI (Fiber Distributed Data Interface) 6.4.5 ATM (Asynchronous Transfer Mode) 6.5 QUESTIONS

50 50 50

7.

IEEE 802.3 - ETHERNET 7.1 The ORIGINS of the ETHERNET 7.2 The COMPONENTS 7.2.1 Transceiver 802.3 7.2.2 Interface 802.3 7.2.3 Drop cable 7.2.4 Repeater 802.3 7.3 ACCESS METHOD 7.3.1 Control of the network access: CSMA/CD 7.3.2 Round Trip Delay 7.3.3 CSMA/CD and MAC sub-level 7.4 PACKET FORMAT at MAC LEVEL 7.5 The PHYSICAL LEVEL 7.5.1 10Base5 7.5.2 10Base2 7.5.3 10BaseT 7.5.4 10BaseF (FB and FL) 7.6 QUESTIONS

54

60

67 68 68 68

71

75 77

82

8.

IEEE 802.3u – FAST ETHERNET 8.1 100BaseTX and 100Base T4 8.2 QUESTIONS

83 83 85

9.

IEEE 802.5 – TOKEN RING 9.1 The ORIGINS of the TOKEN-RING 9.2 The ACCESS METHOD 9.3 TOKEN and FRAME FORMAT 9.4 NETWORK ACCESS CONTROL 9.5 NETWORK CONTROL 9.5.1 Functional addresses 9.5.2 Active Monitor 9.5.3 Control activity 9.6 The PHYSICAL LEVEL 9.6.1 The cabling 9.6.7 The concentrators (MSAU) 9.7 QUESTIONS

86 86 86 87 89 91

92 95

CONTENTS 10. EVOLUTION of LOCAL AREA NETWORKS 10.1 SWITCHED NETWORKS 10.1 Ethernet "switching" 10.2 VIRTUAL LOCAL AREA NETWORKS (VLAN) 10.3 QUESTIONS

97 97 99 100

11. UPPER LEVEL PROTOCOLS 11.1 The TCP/IP PROTOCOL 11.2 The IPX/SPX PROTOCOL 11.3 The NETBIOS/NETBEUI PROTOCOL 11.4 QUESTIONS

101 101 103 105 107

12. CONFIGURATION of a LOCAL AREA NETWORK 12.1 DESIGNING a LAN 12.1.1 Choice of the cabling system 12.1.2 Max. network extension 12.1.3 Ethernet 10Base5 12.1.4 Ethernet 10Base2 12.1.5 Ethernet 10BaseT 12.1.6 Ethernet 10BaseFB/FL 12.1.7 Fast-Ethernet 100BaseTX 12.1.8 Fast-Ethernet 100BaseFX 12.2 MIXED CONFIGURATIONS 12.2.1 The transmission mean and the network technology 12.2.2 The communication protocol 12.2.3 Homogenous grouping of the users 12.3 QUESTIONS

108 108

114

117

13. NETWORK CONTROL AND ADMINISTRATION 13.1 Control and introduction of network information 13.2 Protocols analysis and traffic control 13.3 Statistical analysis 13.4 QUESTIONS

118 118 119 120 121

14. NOS (Network Operating System) 14.1 WHAT IS A NOS 14.2 NOVELL INTRANETWARE 14.3 MICROSOFT WINDOWS NT SERVER 14.4 QUESTIONS

122 122 123 124 125

15. INTERNET / INTRANET NETWORK 15.1 INTRODUCTORY CONCEPTS 15.2 PROTOCOLS and SERVICES 15.2.1 The IP (Internet Protocol) 15.2.2 The TCP (Transmission Control Protocol) 15.2.3 Telnet 15.2.4 FTP (File Transfer Protocol) 15.2.5 TFTP (Trivial File Transfer Protocol) 15.2.6 SMTP (Simple Mail Transfer Protocol) 15.2.7 DNS (Domain Name Server) 15.2.8 SNMP (Simple Network Management Protocol) 15.2.9 NIR (Network Information Retrieval) 15.3 INTRANET NETWORKS 15.4 QUESTIONS

126 126 126

VOLUME 2/2: EXERCISES

129 130

CONTENTS

1. INTRODUCTION to LOCAL AREA NETWORKS

1. INTRODUCTION to LOCAL AREA NETWORKS 1.1 WHAT is a LOCAL AREA NETWORK The main purpose of local area networks in computers (LAN, Local Area Network) is sharing information, and hardware and software resources. Networks of computers are created so that their users can share programs, data and peripheral devices, independently from their physical location. Many definitions of local area network have been attempted. The IEEE (Institute of Electrical and Electronics Engineers, body that has developed an important packet of standards for local area networks) defines the local area network as ... ...a system enabling more independent equipment to communicate directly between them, in the a moderately large geographical area via a physical communication support characterized by a moderate transmission capacity. Let’s examine the single parts of the definition. … The local area network is a system enabling more peripherals to communicate directly between them … A LAN enables an any-to-any communication, i.e. a communication where each equipment can communicate directly to any other equipment connected to the LAN. … The communication occurs within a moderately large geographical area... As already mentioned, this is an important difference between the LAN and the WAN (Wide Area Network). Typically, local area networks are confined inside a single building or group of buildings. … The communication occurs on a physical communication support … In a local area network, the equipment are connected via a private line, which can be a dedicated cable or another communication support. … They use a moderate transmission capacity... Usually, local area networks support moderate transmission capacities: low transmission speeds are used in the WAN and high transmission speeds are used between microprocessors and peripherals. Typical

-1-

1. INTRODUCTION to LOCAL AREA NETWORKS

speeds on the WAN range from 9600/19200b/s to 2Mb/s, while those on the LAN range between 4Mb/s and 100Mb/s. The concept moderate transmission speed formulated by the IEEE today does not fit the reality, as there are local area network technologies supporting communication speeds up to 1Gb/s.

Fig.1.1 Example of local area networks

-2-

1. INTRODUCTION to LOCAL AREA NETWORKS

1.2. ELEMENTS of a LOCAL AREA NETWORK Local area networks generally consist of a set of hardware and software components, including centralized calculation systems, network Server systems, user’s systems, network interfaces, transmission means, concentrators and network software. In general, the LAN can be divided into three macro components: • network technology • network user • network server systems. 1.2.1 Network technology The network technology is the set of systems and physical means which main purpose is the information exchange between different users. It is generally composed of different kinds of physical transmission means or cabling systems, as well as of the network interface cards (NIC) that equip the systems of the local area network. Each network technology is usually identified by its own type and topology. The most used types are those using copper or optical fiber cables as transmission mean. However there are "wireless" devices (via radio) that can cover the same areas without cabling, to the detriment of the transmission speed and safety. The most common cabling topologies consist in the linear bus and tree structures. The bus structures can be used to carry out networks provided with a reduced number of users (usually some tens) and are characterized by low cost and easily achievable cabling. On the other hand, there can be problems in case the trunk should be damaged or cut off, as the functionality of all users connected to it can be compromised. The tree structures need devices that can section the physical trunks, obtaining branches which can be structured in a logical way up to reach the single users. These devices can be simple passive repeaters or intelligent systems, which can classify the information along the single trunk and route them toward the single user, in this way optimizing the network traffic. The most tangible advantages are the capacity to detect and insulate any fault or malfunction on the trunks or the workstations, and the facility in the expansion or reduction of the total structure.

-3-

1. INTRODUCTION to LOCAL AREA NETWORKS

1.2.2 Network users A network user can be a user provided with PC/workstation equipped with particular peripherals/accessories such as printers, hard disks, modems, etc. (obviously, the NIC card included), which can be shared by all components of the system. A network user is usually identified univocally by a proper noun or by particular numerical sequences (logical and physical addresses). The assignment of these parameters is normally remitted to a single figure, called administrator or network supervisor, who has the faculty to create or remove users, establish users unification called groups, their coordinators or any person in charge of them, rights and restrictions on the use of particular network resources. The identification of the single user by means of the name or the numerical sequence will depend on the NOS characteristics (Network Operating System). In any case, to access the available resources, the user must be recognized and identified via two parameters called "username" and "password". 1.2.3 Network servers The network server, commonly called Server, is a computer dedicated to a set of activities which can be classified as follows: • to keep a Database of the users which must be authenticated by the Server. For each single user, this Database will contain all information fixed by the supervisor, such as username, password, rights to resources in the network, time and connection places restrictions, etc. • to control the shared file memorization created by the users, and guarantee safety measures so that integrity and recoverableness are respected also in case of malfunctions • to keep safety copies of the contained files (called backup copies), to be used in case of system’s blocks or for historical data filing and reset. The Server can make any additional peripherals (such as CD-ROM, magneto-optical units, modem, etc.) available, according to the needs of the single users and the rights they are granted by the network supervisor. Another activity assigned to the Server is the control of local messages and electronic mails. Usually, this activity needs the installation of additional software modules. In some situations, more and more common nowadays, the Server can carry out functions of routing and conversions of data formats from and for the network with other systems.

-4-

1. INTRODUCTION to LOCAL AREA NETWORKS

1.3 TRANSMISSION MEANS We already said that, in order to create a network, it is necessary to physically connect the processors and the network equipment by means of proper transmission means. The choice of the type is important as concerns the messages routing techniques, the tolerance to the faults, and the logical organization of the protocols. The most used transmission means in local area networks are the copper lines (twisted pair or coaxial cable) and the optical fibers. Some applications use wireless radio or infrared rays transmission, too. The transmission means and the cabling systems are examined in details in chapter 2.

1.4 NETWORK ARCHITECTURES A network technology is composed of different active and passive elements. The modes with which these elements are interconnected are identified as network architecture or topology. There are different network topologies, which can be divided into the following categories: • bus • star • tree • ring • braid. 1.4.1 Bus topology A bus is the simplest kind of network (fig.1.2). In this topology, all nodes of the network connect directly to the same cable segment. Each node of the network is assigned an address or a number univocally identifying the node. A network segment using a bus topology can be carried out with a trunk of coaxial cable (terminated at the ends), which all users are connected to. When a station on the network transmits a message, the signal travels in both directions until it reaches the cable ends. As the signal propagates on the cable, each connected station can examine the data. The main advantage of the bus topology is its simplicity. The stations can be connected simply by setting a cable from a station to the other. According to the electrical characteristics of a bus, each component on a bus network can affect the whole network. A breakage on the cable or a malfunctioning node can make the whole network unavailable. Just

-5-

1. INTRODUCTION to LOCAL AREA NETWORKS

because it uses a single common transmission mean, the bus topology is not suitable to the structured cable (see chap.2). 1.4.2 Start topology In star topology, each node is connected to a concentration point (Hub) via a separate cable (fig.1.3). The star topology is more expensive than the bus one, as concern the quantity of cable used as for the cost of the necessary concentration equipment. The advantage of this kind of topology lies in the fact that the faults or the malfunctions of a single network user do not affect the operation of the whole network, but only the single trunk between network user and concentrator. In this way, the fault can be insulated and easily detectable in the structure. Usually, a star network is implemented by carrying out "n" point-topoint connections between network users and the concentrator using the multipair copper cable (twisted pair) as transmission mean. 1.4.3 Tree topology It is an evolution of the star topology, as it interconnects more star network users via physical connections that are star cabled, so to create a hierarchical structure called tree (fig.1.4). In the tree, the routing problems are simple as there is a single path to connect two separate users. On the other side, the malfunction of a trunk between two stars insulates their connection. 1.4.4 Braid topology This topology includes redundant connections between different users, via "n" point-to-point channels, creating two or more nested rings (fig.1.5). If all systems are connected between them, we can say it is a complete braid, otherwise it is an incomplete braid. As the structure of a complete braid needs a number of connections increasing exponentially with the number of users, it is better to use it in small networks. The incomplete braid can be applied to geographical networks.

-6-

1. INTRODUCTION to LOCAL AREA NETWORKS

1.4.5 Ring topology In ring topology, each system is connected to the next system using a point-to-point connection, and the last system is connected to the first one (fig.1.6). A one-way ring is created where each system has the repetition function of the message to be transmitted, too. When the system sends a message, it inserts it into the ring and sends it to the next system. All systems repeat the message up to the destination system; this one, repeats the message, receives it and changes a bit in the queue of the same message to confirm the reception. As concerns the physical cabling, the ring topology is not very reliable. A fault or a system off is sufficient to block the whole network. For this reason, cabling are used for ring networks enabling the exclusion of the malfunctioning transmission means from the network (see chap.9, Token-Ring networks).

Fig.1.2 Bus topology

Fig.1.3 Star topology -7-

1. INTRODUCTION to LOCAL AREA NETWORKS

Fig.1.4 Tree topology

Fig.1.5 Braid Topology : a) complete b) incomplete

Fig.1.6 Ring topology

-8-

1. INTRODUCTION to LOCAL AREA NETWORKS

1.5 QUESTIONS

Q1

What is the topology ? 1 2 3 4

Q2

What is the bus in the network topology? 1 2 3 4

Q3

to join two pieces of cable to convert the network protocol to join the connection of a whole floor to connect a terminal to the existent network

Which of the following definitions describes a LAN ? 1

2

3

Q5

The set of conductors taking the data to the network card A linear segment of coaxial cable which the different systems are connected to A set of segments of coaxial cable joined via a concentrator (Hub) A piece of coaxial cable connecting the Hubs to the Switches.

What function does a NIC card perform? 1 2 3 4

Q4

The structural lay-out of the network equipment The physical connection system between terminals The signal coding system The number of NIC cards installed in a terminal

system enabling more independent equipment to communicate directly between them, within a moderately wide geographical area via a physical communication support characterized by a moderate transmission capacity system enabling more independent equipment to communicate directly between them, within a wide geographical area via a physical communication support characterized by a high transmission capacity system enabling more independent equipment to communicate directly between them, within a very restricted geographical area via a physical communication support characterized by a reduced transmission capacity

Which are the typical speeds on the LAN ? 1 2 3 4 5

4-10-16-32 Mb/s 4-10-16-100 Kb/s 2-4-10 Mb/s 4-10-16-100 Mb/s 4-10-16-100 MHz

-9-

2. TRANSMISSION MEANS and CABLING SYSTEMS

2. TRANSMISSION MEANS and CABLING SYSTEMS 2.1 MULTIPAIR cable The basic element of the multipair cable is the Twisted Pair. The twisted pair consists of two insulated copper wires wound one around the other (fig.2.1). The wires pair twisting (pair spiral winding) reduces the electromagnetic interference which the cable is subjected to during the use. Inside the coaxial cable there can be more pairs (4, 25, 50, etc.), each with different pair twisting to reduce the effects of cross-talk. AWG

The twisted pair more used in local area networks installation nowadays has a conductor diameter of 24AWG having a characteristic nominal impedance of 100 Ohm. AWG (American Wire Gauge) is a measurement unit for the diameter used very much in local area networks cables. The diameters and the section of the most used twisted pairs are reported in the following table. AWG

Diameter [mm]

Section [mm2]

22

0.6438

0.3255

23

0.5733

0.2582

24

0.5106

0.2047

25

0.4547

0.1624

26

0.4049

0.1288

The technological improvement provides a continuous increase of the twisted pair pass band that enables transmission speeds of 1000 Mb/s. The twisted pair is the most used cable in local area networks because it represents a low cost universal transmission mean, supporting different signals and services, such as low speed and high speed voice, video, and data. There are different multipair cables: • • •

STP (Shielded Twisted Pair): it is provided with a shield for each pair plus a total shield (fig.2.2) FTP (Foiled Twisted Pair): it is provided with a total screen for all pairs (fig.2.3) UTP (Unshielded Twisted Pair): it is not provided with shield (fig.2.1).

There is a marketing classification dividing multipair cables into categories, according to their performances and as function of the information and services that must transit across them. •

Cat.1: cables for analog telephony - 10 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

• • • •

Cat.2: cables for analog and ISDN digital telephony, or short distance and low speed serial transmissions Cat.3: cables for applications up to 10Mb/s with standard 10BaseT or Token-Ring 4 Mb/s Cat.4: cables for applications up to 16Mb/s with standard TokenRing Cat.5: cables for applications up to 100Mb/s with standards of the last generation, on length under 100 meters.

The category identifies the set of electrical characteristics of the cable such as impedance, diameter, attenuation, cross-talk, etc. As an example, the table of fig.2.4 shows the characteristics of an FTP cable (Belden 1456A), an UTP cable (AT&T 2061) and an STP cable (type-1 IBM). Fig.2.1 UTP multipair cable

Fig.2.2 STP multipair cable

Fig.2.3 FTP multipair cable

Parameter

FTP cable

UTP cable

STP cable

AWG

24

24

22

Characteristic impedance

100 Ω

100 Ω

150 Ω

Attenuation [100 m]

6.6 dB to 10 MHz

6.5 dB to 10 MHz

2.2 dB to 4 MHz

8.3 dB to 16 MHz

8.1 dB to 16 MHz

4.5 dB to 16 MHz

Fig.2.4 Characteristics of FTP, UTP, STP cables

- 11 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.2 COAXIAL CABLE The coaxial cable (fig.2.5) consists of a copper core (hot pole) covered with insulating material (dielectric). This is surrounded by a second layer of conductor material (shield), that can be a braid or a solid element. An insulating protective sheath contains the whole. The coaxial cable is less subject to interference and cross-talk in respect to a twisted pair, and, besides, it can be more easily used for high speed transmissions. The characteristic impedance of the coaxial cable depends on the ratio of the screen sections in respect to the central conductor, and the physical characteristics of the dielectric. The coaxial cables used the most for local area networks are: •

Thick Ethernet: it has an approximated diameter of 10 millimeters and is classified by the code "RG8" (known to the insiders as "yellow cable", as the first time it was introduced in the market it was this color). This cable is used in networks with linear bus topology and connections via transceiver coupled directly to the same cable. The cable is not very flexible, and so it is better to use it in buildings with floating floors. Usually, networks built up with this kind of cable are marked with the code 10Base5, where "10" indicates the speed (10Mb/s), "Base" indicates the base band transmission and "5" the segment length (not over 500 meters).

•

Thin Ethernet: it has an approximated diameter of 7 millimeters and is classified with the code "RG58". This cable is used in linear bus topology networks with direct connection near the NIC network card. Its flexibility makes it good to be used in all types of building installations. Usually, the network carried out with this kind of cable is marked with the code 10Base2, where "10" indicates the speed (10Mb/s), "Base" indicates the base band transmission and "2" the segment length (not over 200 meters, 185 real). The cable RG58 has a higher attenuation than the RG8 (Thick Ethernet).

Fig.2.5 Coaxial cable

- 12 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.3 OPTICAL FIBER The optical fibers can be used to carry data signals as light pulses. An optical fiber consists of a cylinder made with an extremely thin glass, called Core, surrounded by a concentric layer of glass called (Cladding). The light propagates inside the Core by next reflections (fig.2.6), which occur when the light ray meets the Cladding. There is reflection because the refraction index of the Cladding (n2) is lower than the Core one (n1). The actual optical conductor, i.e. the structure consisting of Core and Cladding, is also called naked fiber (fig.2.7). As this structure is mechanically too fragile, it is strengthened by many protection layers, obtaining the single-fiber optical cable. Setting more single-fiber cables together you obtain multiple optical cables (fig.2.8).

Fig.2.6 Light propagation in the fiber

Fig.2.7 Optical fiber structure with layers

Fig.2.8 Multiple optical cable structure

- 13 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

The main advantages of optical fibers are: • high immunity to the electromagnetic disturbances • pass band superior to any copper cable, enabling transmissions with speeds in tens of Gb/s • very low attenuation, in the order of fractions of dB/Km (singlemode fibers) or a few dB/Km (multimode fibers) • reduced dimensions and weight. The disadvantages are the high cost of the cabling system and the network accessory elements. Considering the pros and cons, we can say that the fibers can be used for the construction of backbones or main trunks. The optical fibers can be divided into three categories, as function of the light propagation mode (fig.2.10): • Step-Index and Graded-Index Multimode: these are the optical fibers with more propagation modes. The most used fiber in local area networks is the Graded-Index 62.5/125, where the digits represent the Core (62.5µm) and the Cladding diameter (125µm). In multimode fibers, the light rays propagate in different modes through paths of different length with different propagation times. This causes an enlargement of the received optical pulse, with consequent reduction of the maximum transmission time. The effect of the received pulse enlargement is knows as Modal dispersion (fig.2.9). The Modal Dispersion is affected by the profile of the refraction index inside the Core. In Step-Index fibers (high Modal Dispersion) the refraction index has a step variation, in the Graded-Index fibers (average Modal Dispersion) instead it changes gradually. • Singlemode: these are fibers which core is reduced to 8–10µm, and the fiber behaves as an optical guide having a single propagation mode. This greatly reduces the Modal Dispersion and enables very high transmission speeds (tens of Gb/s). The most used singlemode fiber is the 10/125. Figure 2.11 shows a typical attenuation curve of a singlemode fiber, as function of the wave-length. The optical resources normally used for singlemode fibers are Leds operating at wave-lengths of 850nm (called also Ia Window) or 1330 nm (IIa Window). Photodiodes and PIN photodiodes are used as optical detectors. In singlemode fibers, laser diodes are used as resources (at 1330 or 1550 nm, IInd or IIIrd Window), and Avalanche Photodiodes as detectors.

- 14 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

Fig.2.9 Pulse enlargement due to the Modal dispersion

fig.2.10

Index profile and Modal dispersion a. Step-Index multimode fiber b. Graded-Index fiber c. Step-Index Singlemode fiber

fig.2.11 Typical attenuation curve of singlemode glass fiber

- 15 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.4 WIRELESS The wireless transmission enables to change segments for connection via cable where there are difficulties of physical connection. Electromagnetic waves (radio transmission) or signals in the infrared range are used. The wireless transmission has also disadvantages. The radio transmissions are subject to interference, which can cause transmission errors. The coverable distances are limited. The non radio transmissions, e.g. infrared , need the absence of physical barriers between transmitter and receiver. The pass band of these systems is usually lower than cabled systems. 2.5 CABLING SYSTEMS A technological plant for information transmission in all its different shapes (local area networks, video images, telephony, etc.), is called cabling system. The rules on the cabling systems define the methods to cable a group of buildings constructed in an interbuilding. The rules describe: • the characteristics of the transmission means and the passive components (connectors, terminators, user’s sockets, adapters, etc.) in relation to the required transmission speeds; • the allowed cabling topologies (star, bus, ring, etc.) and their characteristics (maximum distances, etc.); • installation rules. 2.5.1 Proprietary cabling The need of cabling systems for information units appeared at the beginning of the 80s from the growing need to connect electronic equipment. In those years, the first local area networks (Ethernet and Token-Ring) were created, and consequently, the first proprietary cabling systems were developed, e.g. IBM "Cabling System" and Digital "DECconnect". Each of these cabling systems consisted of a set of proprietary elements (cables, connections, cross-connection panels, etc.) which, although satisfying the "cabling" needs of that period, could hardly be integrated between them due to the intrinsic difference of the solutions suggested by the different manufacturers.

- 16 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.5.2 Structured cabling The evolution of local area networks caused the proprietary transmission means to be abandoned, and favored the systematic use of the copper twisted pair and the star topology. This choice needed the creation of reference rules. So, the standard EIA/TIA 568 was developed, which specifies the minimum requisites for cabling a building or a group of buildings belonging to the same interbuilding. Today, there are the following main standards for cabling systems: • EIA/TIA 568: American standard for cabling commercial buildings. Approved in 1991, it is the most used nowadays • EIA/TIA 570: American standard for cabling residential buildings, which can have a small number of commercial offices • ISO/IEC DIS 11801: proposal of international standard for cabling of commercial buildings, approved in 1994. This standard is always asked as base requirement to carry out structured cabling • SP-2840-A: proposal for revision of the standard EIA/TIA 568 to face the needs of faster cabling speeds. Approved in 1995, it is called EIA/TIA-568/A • prEN 50173: proposal for European standard, not approved yet and much similar to ISO/IEC DIS 11801. The aspects related to the mechanical and building technology are treated in the standard EIA/TIA 569. The standard TIA/EIA 607 deals with the problem of creating a grounding plant that can fit the structured cabling. 2.6 The STANDARD EIA/TIA 568 This standard specifies the minimum requirements for cabling a building or a group of buildings belonging to the same interbuilding (fig.2.12). The limits of the interbuilding are: • the maximum geographical extension is 3.000 m 2 • the maximum surface of the buildings is 1.000.000 m • the maximum population of the building is 50.000 persons. The cabling must provide a proper support to different communication equipment (voice, data, video,..), and so it must be independent from them. The technical specifications of the standard are: • the topology • the elements belonging to cabling • the transmission means • the backbones • the horizontal cabling • the installation rules • the cables identification • the documents.

- 17 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

Fig.2.12 Model EIA/TIA 568

Fig.2.13 Star topology

- 18 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.6.1 The topology The cabling topology is star hierarchical kind. Other topologies, such as the bus and ring one, typical of some standards for LAN, must consequently be taken back to star topology (fig.2.13). 2.6.2 Cabling elements The main elements constituting a cabling system, as in the EIA/TIA568 definitions, are described hereafter. Some elements (Work Area, Equipment Room, Interbuilding Entrance Facility, Private Branch Exchange) are mentioned in the standard, but are not subject to specifications. •

Main Cross-connect (MC): it identifies a technological place or a distribution closet, located in the central or main building of an interbuilding, distributing the backbone cables to the other buildings. It is the first level of cabling hierarchy (star center of interbuilding or of single building)

•

Intermediate Cross-connect (IC): it identifies the technological place or the distribution closet of a building belonging to an interbuilding, distributing the backbone cables to more-floor buildings. It is the second level of cabling hierarchy (star center of a building). When a cabling is carried out on a single building, the technological place or the distribution closet of a building becomes the first hierarchical level and so is considered Main Cross-connect

•

Telecommunication closet (Telecommunication Closet - TC): it identifies the telecommunication closet distributing the cables (horizontal cabling) reaching the user (fig.2.14). It is the third level of cabling hierarchy (star center of a floor).

•

Interbuilding Backbone: backbone for interconnection between the interbuilding star center building and another building. It starts from the Main Cross-connect and ends on an Intermediate Cross-connect

•

Interbuilding Backbone: building backbone, it is the backbone for interconnection between the building technological room and the telecommunication closet.

•

Equipment Room (ER): it is a technological place which can contain passive equipment, such as cross-connection panels, voltage dischargers, channels and fairleads, and can house active equipment such as the private branch exchange, the LAN concentrators and the audio and video equipment. Its function is very similar to a group of distribution closets, but the larger dimensions available make it good to be used as interbuilding or building star center.

- 19 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

•

Interbuilding Entrance Facility (EF): set of technologies and passive components used for the entrance of the building interbuilding backbones. In the EF, electrical protections are required for the copper cables, and the grounding of the different components must also be considered

•

Transition Point (TP): it is the transition point (in horizontal cabling) between a twisted cable and a flat cable. This possibility must be used only for low frequency transmission (ten of kHz), as the flat cable has a very high cross-talk

•

Cross-connect ( or Distributor): it is composed of at least two crossconnection panels, one for the incoming cables and the other for the outgoing ones. The mechanical termination of the copper cables is made on termination blocks called also "wiring blocks". The termination of the optical fiber cables is carried out on proper panels

•

Patch panel: cross-connection unit for the transmission means (fig.2.15). It can take two shapes: − for copper cables, it can contain one or more termination blocks − for optical fibers, it can contain a set of passing connectors, (called barrels or compasses), used to cross-connect the fibers between different panels or between a panel and an active equipment.

•

Patch Cord: patch cords for copper cables or for optical fibers.

•

Telecommunication Outlet (TO): outlet which can contain one or more connectors (fig.2.16)

•

Adapter: it is a cabling adapter, and the standard states that it must be installed externally to the users telecommunication outlets. It can be: − passive, to adapt different connectors or cables − active, to adapt different transmission systems, such as RS232RS422 converters, modems, etc.

•

Work Area (WA): it identifies the user’s work area or desk

•

Private branch exchange - PBX): it identifies the private branch exchange, controlling the voice communications

- 20 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

Fig.2.14 Telecommunication closet

Fig.2.15 Patch Panels

Fig.2.16 Telecommunication Outlets

- 21 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.6.3 The transmission means The following transmission means are allowed: • 50-Ohm coaxial cables • 62.5/125-µm singlemode optical fibers • UTP cables with 4 100/120/150 Ohm pairs • STP cables with 4 100/120/150 Ohm pairs. The characteristics required for the coaxial cables are those specified by the standards IEEE 802.3, 10Base5 and 10Base2 (see next chapters). The allowed optical fiber is the multimode one with 62.5/125µm. Each single conductor of the 4-pair UTP cables is 24AWG. They must satisfy at least the characteristics of the category 3. The pairs are identified by the following colors: • Pair 1: white-blue (W-BL) and blue (BL) • Pair 2: white-orange (W-O) and orange (O) • Pair 3: white-green (W-G) and green (G) • Pair 4: white-brown (W-BR) and brown (BR). The multipair UTP cables contain one or more groups each with 25 pairs, the conductors have a dimension of 24AWG, but 22AWGconductors are accepted, too, only if respecting the minimum required electrical characteristics. 2.6.4 The backbones The Backbones are the cabling carrier elements and can interconnect, with star hierarchical topology: • different buildings to the star center of the interbuilding (Interbuilding Backbone) • Different floor closets to the building closet (Interbuilding Backbone). The following cables are allowed: • 100/120/150-Ohm multipair UTP wires • 100/120/150-Ohm STP multipair STP cables • 62.5/125-µm multimode optical fiber • 50-Ohm coaxial cable, type Thick Ethernet, addressed to the two ends with "N" connectors.

- 22 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.6.5 The horizontal cabling The Horizontal Cabling interconnects the different work areas in the telecommunications closet, and must provide at least the following services: • voice transport • serial data transmission • data transport for local area networks • signal transport for the control of devices inside the building (thermostats, alarms, …). The topology is star starting from the telecommunication closet. The maximum allowed distances for the distribution cables and the patch cords are indicated in fig.2.17. The following cables are allowed: • 4-pair UTP cable with 100-Ohm impedance • 2-pair STP cable with 150-Ohm impedance • 50-Ohm coaxial cable, type Thin-Ethernet, addresses to the ends with BNC connectors • 62.5/125-µm singlemode optical fiber. The outlet related to the single work area must contain at least two cables, of which at least one must be 4-pair UTP, of category 3 or superior. The UTP cable must be addressed to an outlet RJ45 (fig.2.18). The second cable must be any of the listed cables allowed for horizontal cabling.

Fig.2.17 Horizontal cabling

- 23 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.6.6 Cables identification To ease the control and maintenance of cabling systems, it is necessary to use standard rules for cables identification. The standard states that the backbone cables must have a single number, which must contain at least two ranges indicating: • the cable identification • the number of pairs, in case of multipair cable, or the number of fibers in case of multi-fiber cables. A numbering example of a backbone cable is the following:

4005/1-300 indicating a cable with number 4005 and containing the pairs from 1 to 300. Each work area and the related cable are identified by a label, usually composed of 8-10 characters that may include numbers or alphabetical letters. The numbering must contain: • the reference to the floor of the building where the work area is located • the reference to the telecommunication closet to which the work area is connected • a field of three characters identifying the same work area. Usually the floor closets are identified by alphabetical letters. An example of how the work area is numbered and the related cable is shown in the label:

EVB02-16G having the following meaning: • • • •

EVB indicates the name of the building: Elettronica Veneta Building 02 indicates the floor of the work area 16 identifies of the work area G identifies the telecommunication closet the work area is connected to.

2.6.7 Documents Each cabling must be provided with documents written out with standard symbols and abbreviations. It must include: • The logical drawing inside the whole interbuilding or the single building • a table to identify the backbones

- 24 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

•

a closet table indicating the connections telecommunication closet and the work areas.

between

the

The backbone documentation table must contain: • the identifications of all cables and their corresponding number of pairs or fibers • the localization and the identification of the two closets each cable is attested to. Each telecommunication closet must contain its documentation in a proper place. This documentation consist in cross-connection cables, with which it is possible to reconstruct the path of the cable that, starting from a particular position of the cross-connect, reaches the work area: the active pairs and their use must also be indicated. 2.6.8 Types of connectors and splicing systems The following connectors are allowed: • RJ45 connector: for 4-pair UTP cables (fig.2.18) • Hermaphrodite IDC connector (IBM Data Connector): for 2-pair STP cables (IBM cables for Token-Ring networks) (fig.2.19) • "N" connector: for backbone coaxial cables • "BNC" connector: for horizontal distribution coaxial cables • optical fiber connectors: they must be able to support at least 200 cycles of extraction/insertion without introducing attenuation over 1 dB. The most used ones are the "ST" and the "SC" (fig.2.20) • mechanical optical junctions (splices): they are used to join two trunks of optical fiber. The maximum allowed attenuation on the junction is 0.3dB.

fig.2.18 RJ45 connector a) female b) male

fig.2.19 IDC connector

fig.2.20 a) ST connector b) SC connector

- 25 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

2.7 QUESTIONS

Q1

With reference to the transmission means, what does the twisted pair consist of? 1 2 3 4

Q2

What is a category in a twisted pair? 1 2 3 4

Q3

Impedance Length Cost Diameter

Which of the following advantages characterizes the optical fiber cables ? 1 2 3 4

Q5

The number of wires the twisted pair consists of The classification of the single insulator The electrical characteristics of the cable as function of the cabling type The price for 100m of cable

Which is the most important electrical parameter characterizing a coaxial cable? 1 2 3 4

Q4

Pairs of twisted coaxial conductors Pairs of twisted insulated conductors Double pairs of twisted copper conductors Single wires contained into an insulating sheath

Easy connection Pass band Weight Cost

Which bodies defined the standard for structured cabling? 1 2 3 4

E.N.E.L. ANSI EIA ASCII

- 26 -

2. TRANSMISSION MEANS and CABLING SYSTEMS

Q6

What is a topology ? 1 2 3 4

Q7

What does UTP define in a twisted pair? 1 2 3 4

Q8

The structural lay-out of the network equipment The physical connection system between terminals The signal coding system The number of NIC cards installed in a terminal

The presence of shielding The absence of shielding A connector for twisted pair cables The equipment it is connected to

How must the cables constituting a network cabling be identified? 1 2 3 4

It is not necessary to identify them With the color Via numbering chosen by the cabler Via numbering identifying position and work area

- 27 -

3. NETWORK ACTIVE COMPONENTS

3. NETWORK ACTIVE COMPONENTS In designing a local area network, once the network architecture is defined, the active equipment must be chosen necessary to carry out the LAN. The typical used components are the network card (NIC), the Transceiver, the Hub, the Repeater, the Bridge, the Switch and the Router. 3.1 Network Interface Card (NIC) The network interface card is installed in all devices directly connected to the network. The card provides all hardware functions necessary to make the calculation equipment (Personal Computer, Host, Server, Workstation, ..) accessible to the communication means (LAN). The card must be compatible to the existent cabling. The Ethernet cards have usually a BNC connector (for RG58 coaxial cable - Thin Ethernet) or a RJ45 connector (UTP cable), and almost all have a port called AUI (Attachment Unit Interface) for the connection to the Ethernet transceiver type Thick (yellow cable). The cards provided with more connectors are called combo. According to the Bus of the PC/workstation on which they are mounted to, it is possible to classify them in terms of speed. The cards for ISA bus are the less powerful, they transfer 16-bit data and operate at 8 MHz, while those for PCI, EISA or MCA bus transfer the 32-bit data with frequencies ranging between 25 and 33 MHz. Another important characteristic concerns the way in which the cards are controlled by the CPU of the PC/workstation. The ISA cards can be controlled in I/O or DMA, but do not allow the "bus-mastering" technique which enables a fast exchange of information between the network card and the RAM of the system, with no help from the CPU. This technique is native on PCI and EISA buses. Each card inserted into a bus must be recognized by the CPU from a particular set of hexadecimal addresses (called I/O). With these addresses the driver software can initialize the card controller. The addresses are usually included inside the following ranges: 300...31FHEX 320...33FHEX 340...35FHEX 360...37FHEX Other values can be suggested by the different manufacturers. Remember that the address 300H would be not proper as dedicated to experimental cards. Even the Interrupt lines (IRQ) available on a system of AT-compatible class are not many: INT 5 INT 10 INT 11 INT 12 INT 15 - 28 -

3. NETWORK ACTIVE COMPONENTS

We suggest an accurate choice, to prevent conflicts with other peripherals with consequent malfunctioning or possible disconnection. The Hardware configuration of the NIC card, i.e. the choice the address and the IRQ line, must always be carried out before installing the software driver. The card can be configured in one of the following ways: • via jumpers • with configuration programs memorizing in EEPROM (of the NIC card) the user’s setting. The devices in a NIC card are many. We report the main ones: • Microprocessor: it synchronizes and controls the operations developed by the communication controller, optimizing the data flows of the buffers of the local RAM and the system bus. Actually there are cards integrating microprocessor and LAN controller into a single integrated circuit, in this way reducing the costs • Network Controller: it controls the whole communication from and toward the card, it constitutes the data packets and calculates the CRC, it recovers errors and activates eventual retry procedures • Local RAM: it represents the card transmission and reception buffers. Usually it ranges between 16 and 64KB • EEPROM: it memorizes the user configurations also without power supply voltage • Transceiver: it is the matching circuit for input and output analog signals from the card, toward the transmission mean. It is chosen by the manufacturer according to the network connection supported by the card (10Base2, 10BaseT, 100BaseT, etc). In the 10Base5 connections, the Transceiver is external and is connected to the card via a cable which maximum length is 50m called drop-cable or cable AUI/AUI. The purpose of the last is to power the external transceiver (see next chapter).

Fig.3.1 Network card

- 29 -

3. NETWORK ACTIVE COMPONENTS

3.2 TRANSCEIVER The Transceiver is the element enabling the transmission/reception of the packets between the AUI interface of the NIC card (Ethernet controller) and the transmission mean. This component typically used in Ethernet networks is carried out with yellow coaxial cable (Thick Ethernet), but it is often employed also to convert the format of the signals between the AUI interface and the other transmission means (twisted pair, thin coaxial cable, optical fiber). The NIC card is connected to the Transceiver via a drop-cable. The transceiver consists of: • a transmitter sending data from the interface to the transmission mean • a circuit for collision detection (see chapt.7), sending the interface (card) a collision signal in case this is occurred. The transceiver can also carry out a test on the proper operation of the collision detector, and send the result to the interface. This test can also be enabled or disabled (on the Transceiver) and can take different names: Collision Presence Test (CPT), Heartbeat (HB), Signal Quality Error_test (SQE) • a receiver for the data coming from the transmission mean and going to the interface of the NIC card. In case of 10Base5 networks (yellow cable, Thick Ethernet), the Transceiver is addressed to the Thick cable with a complex mechanical mounting operation. The external part breaks the cable sheath and enters in contact with the copper shielding constituting the second pole of the coaxial cable. The inner core of the cable is reached by a needle, which must penetrate across the shielding without touching it not to cause a short-circuit. The connection is called vampire connection. A variation includes the use of a Transceiver with "N" connector, which does not need particular mounting operations as it has a reliable quick acting connector.

Fig.3.2 10Base5 Transceiver

Fig.3.3 Transceivers a) AUI-10Base2 b) AUI-10BaseT c) AUI-10BaseF

- 30 -

3. NETWORK ACTIVE COMPONENTS

3.3 HUB The star topology of the local area networks includes the use of concentrators (Hubs), which enable more equipment to be connected to the cabling system of the LAN across a single concentration point. This device usually has an input port selectable between coaxial cable, optical fiber or twisted pair, and a particular number of equal ports (4, 8 16, 32, etc.). The use of the Hub enables to create star topologies, which center is the same Hub. Each port can be active or sectioned as function of the state of the communication channel. This state is displayed via single light indicators for the ports. In some models, there are collision and traffic indicators and speed selectors. Note that each port of the Hub can be used for a point-to-point connection with the PC/workstation or a possible network Server. The Hub operates as Repeater (see next chapter), operates at level 1 (Physical ) of the OSI model (see chapt.6), and is transparent to all protocols operating at higher levels.

Fig.3.4 Use of the Hub

Fig.3.5 Hub

- 31 -

3. NETWORK ACTIVE COMPONENTS

3.4 REPEATER The Repeater enables to connect single network elements in order to create a wider network. The function of a Repeater is to receive a message and to re-transmit it, regenerating the original signal intensity. A LAN usually has a restriction on the physical size of any single element of the network, determined by the transmission mean and the transmission technique. The Repeater enables to overcome this restriction. The main characteristic of a Repeater lies in the fact that the signals generated on the first segment of cable are propagated to the second segment in a transparent way, without any filtering. They are simple to use and have a contained cost. To use a Repeater, both network segments must be the same kind. All stations must have single addresses. The Repeater operates at level 1 (Physical) of the OSI model (see chapt.6), and is transparent to all protocols operating at higher levels.

Fig.3.6 Bus topology with more segments connected via repeater

Fig.3.7 The Repeater operates at Physical level

- 32 -

3. NETWORK ACTIVE COMPONENTS

3.5 BRIDGE The Bridge enables to connect two separate local area networks and enables the users of each network to access the resources of the other, similarly to the Repeater but in a more intelligent way. The Bridge can implement a filtering mechanism. It receives all frames transmitted by the different network segments to which it is physically connected (usually 2) and then, according to the destination address of each frame, it determines if the last must be transmitted to the other segment. To use a Bridge, both network segments must be the same kind. In this case, too, all stations must have single addresses. A pair of Bridges including telecommunication services can be used to interconnect two separate local area networks, one far from the other. A Bridge operates at level 2 (Data Link) of the OSI model (see chapt.6) and is transparent to all protocols operating at higher levels.

Fig.3.8 More networks connected via Bridge

Fig.3.9 The Bridge operates at Data Link level

- 33 -

3. NETWORK ACTIVE COMPONENTS

3.6 SWITCH The Switch is an active networking device which can maintain more simultaneous connections between stations, at the maximum transmission speed allowed by the local area network. It is typically employed when, inside a network, two or more stations need a direct connection between them for the most time, due to the intense data traffic they generate. In these cases, it is just the Switch that fixes a virtual channel dedicated to the stations generating the traffic, reducing the collisions on all packets of the other stations and actually reducing the network traffic. The Switches can be programmed as function of the needs of the single users, so to privilege any user needing larger band (data exchange volume, expressed in bits per second) in the network. Unlike the Repeater and the Bridge, it is possible to connect networks using different transmission techniques (LAN Ethernet, FDDI or TokenRing). A Switch usually operates at level 2 (Data Link) of the OSI model (see chapt.6), and is transparent to all protocols operating at higher levels.

Fig.3.10 Use of a Switch

- 34 -

3. NETWORK ACTIVE COMPONENTS

3.7 ROUTER The Routers enable to route the messages from a system to another in case of different interconnection paths. The Routers are more intelligent than the Bridges, and can be used to construct highly complex business networks. The function of the Router is to filter the packets generated by a group of users and route them toward other groups as function of the single work stations, each time choosing the best, quicker and more reliable path. In respect to the Bridge, the Router needs a more detailed knowledge of the protocols used to carry the messages. The Router can advantageously be used when there are separate groups organized according to the function (administration, storehouse, production, etc.) or gravitating on separate LANs, or when the LAN are connected across geographical networks (WAN). In these situations, the Router controls the traffic packet in a group or from a group to the other, preventing interference between the packets which destination belongs to the same group and those of another one, and at the same time they let those addressed to different groups pass. The Routers are completely programmable from stations inserted into the network, or via serial ports present on the same devices. Unlike the Repeater and the Bridge, it is possible to connect networks using different transmission techniques (LAN Ethernet, FDDI or Token-Ring). A Router operates at level 3 (Network) of the OSI model (see chapt.6) and is transparent to all protocols operating at higher levels.

Fig.3.11 Networks connected via Router

Fig.3.12 The Router operates at network level

- 35 -

3. NETWORK ACTIVE COMPONENTS

3.8 QUESTIONS

Q1

Which function does a NIC card develop? 1 2 3 4

Q2

Which main function is developed by a hub ? 1 2 3 4

Q3

3 4

to convert a digital into analog signal to exchange the electrical connections between two connectors for UTP twisted pair between them to re-calculate the checksum (CRC) of the damaged packets in order to make them valid to route the data packets making the connections more flexible and optimizing the network traffic

Which component is normally used to interconnect two separate LANs of the same kind, filtering the data passing from a network to the other? 1 2 3 4 5

Q5

to centralize more remote connections in a point to convert a synchronous into asynchronous system to connect networks located in different floors or buildings dividing the traffic to authenticate a user enabling its access to the network.

Which is the main purpose developed by the router ? 1 2

Q4

to join two trunks of cable to change the network protocol to join the connection of a whole floor to connect a terminal to the existing network

Router Repeater Hub Bridge Switch

Which of the last components is normally used to increase the network dimensions?

- 36 -

4. SIGNAL CODING TECHNIQUES

4. SIGNAL CODING TECHNIQUES

4.1 INFORMATION TRANSMISSION To carry out a local area network you must physically connect the processors and the equipment via proper transmission means. The digital transmission mainly requires the coding of the information (the bits) into a format that can be carried by the transmission mean, so that the same information can be properly reproduced in the reception equipment. In the digital systems, the data bits are represented by electrical signals. The simplest form uses two levels to represent the binary digits 0 and 1, e.g. +5 for 1 and 0V for 0. Usually a level is kept fixed for the duration of a bit, and so in this case we speak of NRZ format (Non Return-toZero) (fig.4.1), too. The signal is transmitted across a communication channel, having limited bandwidth, and in reception it is read in particular moments to determine if the coming data is 0 or 1. Before transmission, the data provided by the source are processed (or coded), in order to generate a signal with the following main requisites: • it has a spectrum matching the transmission channel, to exploit the pass band of the same channel at best • it facilitates the extraction of the reception bit timing (clock), to prevent the proper reading of the received bits • it reduces the effects of cross-talk on the transmission cables nearby • it minimizes the electromagnet emissions. The coding acts on two aspects of the signal: • it changes the shape of the symbol associated to the information to be transmitted (see the Manchester coding and MLT-3) • it changes (according to particular rules) the sequence of bits to be transmitted (see, e.g. the 4B5B and 5B6B coding). This chapter describes the main signal coding techniques used in Local Area Networks.

- 37 -

4. SIGNAL CODING TECHNIQUES

Fig.4.1 Signal coding forms

Fig.4.2 Spectrum of NRZ and Manchester signal

- 38 -

4. SIGNAL CODING TECHNIQUES

4.2 NRZ (Non Return to Zero) and RZ (Return to Zero) CODING The simplest kind of coding uses two different voltage levels to represent the two binary digits 0 and 1. E.g. +V for bit 1 and 0V for bit 0 (NRZ unipolar), or +V and -V (NRZ polar, fig.4.1). The NRZ code is generally used to generate or interpret digital signals: PCM coders in telephony or equipment in computer science environment (computers, printers, video terminals, etc.) provide and receive data in NRZ format. Most of the energy of an NRZ signal is concentrated between the d.c. component and half the transmission speed Fb. If, e.g. the transmission speed is 10 Mbit/s, the spectral components will practically result concentrated between 0 and 5 MHz (fig.4.2). The main restrictions of the NRZ signal are: • the presence of a d.c. component in the spectrum, which makes the signal not good to be transmitted • the difficulty to ensure a proper synchronism extraction by the receiver. Consider that in presence of long sequences of bits 0 or 1 the signal has a constant voltage along the whole time period, and this complicates the synchronism extraction in the receiving circuits further. This problem can be reduced associating a stable value to each bit "1" for half bit time. This coding is called RZ (Return to Zero). Due to these problems, the NRZ format, used to generated or interpret digital signals, is not employed (alone) for their transmission. In some applications, the data are transmitted in NRZ format, but the sequence of bits to be transmitted is coded before, always to guarantee a particular minimum number of transitions on the physical mean. Examples of these coding are 4B5B and the 5B6B, described after. 4.3 NRZI (Non Return to Zero Inverted) CODING The NRZI coding (together with the 4B5B) is used for FDDI networks at 125 Mb/s. The NRZI coding inserts a transition (in the center of the bit interval) for the bits “1”, and no transition for the bits “0” (fig.4.1). The transition can be from high to low level or from low to high level according to the state of the signal in correspondence of the last bit. The coded signal shows transitions in the center of the bit interval only if the datum is “1”. The NRZI is also known as NRZ Differential, and the coding law can be explained as follows: the coder inverts the output bit (n+1) if the input bit (n) is "1", it keeps the output the same if the input bit (n) is "0". In

- 39 -

4. SIGNAL CODING TECHNIQUES

other words, the bit “1” is coded as variation of the output bit, the bit “0” as no variation. To synchronize the receiver also in presence of long sequences of zeroes, in case of the FDDI we appeal to a preventive 4B5B coding of the data flow. 4.4 MANCHESTER and DIFFERENTIAL MANCHESTER CODING The Manchester coding is used in Ethernet – IEEE802.3 networks (at 10 Mb/s). The Token Ring – IEEE802.5 networks (at 16 Mb/s) use the Differential Manchester coding. The Manchester coding (known also as Biphase) inserts a transition in the center of the bit interval. If the NRZ datum is 1 there will be a transition from high to low level, if it is 0 a transition from low to high level. In practice there is the transmission of a negated clock period if the bit is 1, of a direct clock period if the bit is 0 (fig.4.1). The spectrum of the Manchester signal points out the presence of a component with frequency equal to the bit speed (fig.4.2), and this facilitates the reception timing extraction. See that the spectrum of the Manchester signal is wider than the NRZ one. This requires a channel with double band in respect to the one necessary without coding. The simplicity makes the Manchester convenient in those applications (as the Ethernet network at 10 Mb/s) where the considerations on the band are not essential. One of the restrictions of Manchester coding is that the signal for the bit 1 is exactly the reverse of the signal for the 0. In many cases, the transmission mean can make difficult or impossible the determination of the absolute polarity or of an absolute phase reference. Consider, e.g. the wires inversion in a twisted pair. In these cases all the 1 transmitted would be received as 0 , and the 0 as 1. The Differential Manchester code (fig.4.1) introduces a variation at the beginning of the bit interval if the datum is “0”, no variation if the datum is “1”. The bit is always represented with a clock period which, according to the mentioned rules, can be direct or inverted. The Differential Manchester is also known as Biphase Differential (in the two variations Mark and Space), and the coding law can be explained as follows: • in Biphase Space coding each bit interval starts with a transition (from high to low or vice versa). Besides, if the datum is 0 there is a second transition at the center of the bit interval, while if the datum is 1 there is no transition • the Biphase Mark is essentially the same, with the difference that the transition at the center of the bit interval is introduced when the datum is 1. - 40 -

4. SIGNAL CODING TECHNIQUES

As a result of Differential coding there is a wave-form in which a binary digit is coded with a clock period, the other digit with a fixed level (alternatively high and low) for the whole duration of the bit interval. 4.5 MLT-3 CODING The MLT-3 coding produces a signal at three voltage levels (+, 0, -), instead of the two NRZ levels. The coding law is pointed out in fig.4.3, where you can see there is always a transition at the center of bit 1, and no transition for the bits 0. The transitions for the bits 1 go in the following order: 0 → + , + → 0 , 0 → - , - → 0 , etc. Note that, in the worst case of transmission of all bits “1”, the main frequency of the transmitted signal reduced to ¼ of the bit speed. The MLT-3 coding is used by FDDI TP-PMD and Ethernet IEEE 802.3 100BaseTX, two standards for 100-Mb/s transmissions on copper cables. For FDDI and Ethernet, the transmission speed on the transmission mean is 125 Mb/s, and so the main frequency is 31.25 MHz. The difference of 25 Mb/s between the nominal speed at Data Link level (100 Mb/s) the physical level speed (125 Mb/s) is due to the fact that the absence of transitions for sequences of bits 0 forces, in this case, too, a pre-coding type 4B5B of the sequences to be transmitted (fig.4.4). MLT-3 has been also proposed for the ATM at 155 Mb/s on copper, with 4B5B coding, so the frequency of the fundamental is 48.4375 MHz.

Fig.4.3 MLT-3 coding

- 41 -

4. SIGNAL CODING TECHNIQUES

4.6 4B5B - 5B6B – 8B6T CODING For all coding techniques analyzed, except the Manchester one, there are data sequences which do no generate transition. To guarantee the transmission of a number of transitions sufficient to enable the receiver synchronization, it is necessary to pre-code the data to be transmitted, eventually enlarging the sequence. In local area networks, this function is developed by two codes: • the 4B5B code (4_Binary-to-5_Binary), coding each possible sequence of four bits into five bits and is used together with NRZI or MLT-3 (fig.4.4) • the 5B6B code (5_Binary-to-6_Binary), transforming the sets of five bits into sequences of six and is used in the standard 802.12 together with NRZ coding (fig.4.5). The 8B6T code (8_Binary-to-6_Ternary) is used in Ethernet 802.3 100BaseT4 to convert sets of eight bits into group of six ternary symbols. It is a single coding preventing the double passage 4B5B and then MLT-3 of FDDI TP-PMD.

Fig.4.4 4B5B coding

Fig.4.5 5B6B coding

- 42 -

4. SIGNAL CODING TECHNIQUES

4.7 QUESTIONS

Q1

Which of the following definitions explains the electrical signal attenuation? 1 2 3 4

Q2

Which measurement unit is it used to measure the transmission speed? 1 2 3 4

Q3

three bits have duration equal to the main frequency period a bit has a duration equal to half the main frequency eight bits have duration equal to the main frequency period six bits have duration equal to 5/6 the main frequency period

How are the bits coded in Manchester coding? 1 2 3 4

Q5

Megabit per second MegaHertz per second Decibel per second Volt per meter

How are the signal bits coded according to the NRZ standard? 1 2 3 4

Q4

its maximum frequency the number of odd harmonics in the signal the reduction of the signal amplitude crossing a fixed length of cable the modification of the signal wave-form.

a period of the direct clock signal for the zero, inverted for one a period of the direct clock signal for one, inverted for the zero two periods of the clock signals if the bit keeps constant, a period of the clock signal if the bit changes two periods of the clock signal if the bit changes, a period of the clock signal if the bit keeps constant

How many logical levels are used to code a differential Manchester signal ? 1 2 3 4

Three Two One Four

- 43 -

4. SIGNAL CODING TECHNIQUES

Q6

Which are the advantage of the MLT-3 coding ? 1 2 3 4

Q7

none, it is simply a different coding standard increasing the number of levels you reduce the frequency of the fundamental with the same speed more channels can be transmitted on the same cable cables of inferior quality can be used

What is the 4B5B in signal coding systems? 1 2 3 4

4 bits divided on 5 terminals the coding of 4 bits with 5 bits in line the coding of 4 bytes each 5 clock periods the coding of 4 bytes with 5 transitions of the line signal.

- 44 -

5. ACCESSES CONTROL TECHNIQUES

5. ACCESSES CONTROL TECHNIQUES 5.1 ACCESS METHODS One of the main aspects characterizing a LAN is the access method, i.e. the basic modes in which a station connected to the network can access the transmission mean, transmit the available data and give the physical line to another station. Integrating parts of the access method are also the other characteristics, such as: • the time in which a station can appropriate the transmission mean for its exclusive use • the time after which it is forced to disconnect and give this right to the other stations • the modes enabling to prevent more stations to simultaneously transmit the available data (making them mutually incomprehensible to their destinations). There are different access methods, which can be inserted into one of the following categories: •

deterministic access: the access modes to the bus belonging to this family enable the station to transmit in fixed instants and with rules of behavior accurately specifying: − the priority of a station − the maximum transmission time − how to transfer the right to transmit to the other station of the local area network. One of the networks using this method is the Token-Ring

•

non deterministic access: it gives a high freedom to a station which is to transmit data. The transmission does not occur at fixed or established times, and usually there is no priority level. Enabling a simple implementation of the transmission protocol, these kind of methods are provided with reliable mechanisms for collision detection (i.e. in cases in which the two stations transmit simultaneously), and transmission repetition, using particular algorithms to prevent the continuous repetition of the same collisions.

- 45 -

5. ACCESSES CONTROL TECHNIQUES

5.2 Carrier Sense Multiple Access / Collision Detection (CSMA/CD) In bus, star and tree topology networks, the most used access control methods is the CSMA/CD (Carrier Sense Multiple Access/Collision Detection), included in the family of non deterministic access methods. Before transmitting, a station listens to the transmission mean to determine if another station is transmitting at that moment (fig.5.1). If no one is transmitting so the station sends its message. The term Carrier Sense (carrier detection) indicates the listening before transmission. Once transmitted, the message runs toward all stations of the network. Each station receives it, examines it and, if it belongs to it, processes it. Sometimes two or more stations can send their messages simultaneously causing a collision (fig.5.2). The transmitting stations stop the transmission and set to pause for a fixed time before transmitting again. Each station generates a casual number determining the stand-by time. The algorithm used for the generation of this number is designed to minimize the stand-by time when the traffic is light and to reduce the number of next collisions when the traffic is heavy. The CSMA/CD method is advantageous when the propagation delay between source and destination is short in respect to the transmission time of the packet. This access method is much used, and its most famous application is in Ethernet-IEEE802.3 networks (see chapt.7.3).

Fig.5.1 CSMA/CD transmission without collisions

- 46 -

5. ACCESSES CONTROL TECHNIQUES

Fig.5.2 CSMA/CD transmission with collisions

Fig.5.3 Token Passing

Fig.5.4 Token Bus

- 47 -

5. ACCESSES CONTROL TECHNIQUES

5.3 TOKEN-PASSING In ring topology networks, the most used access control method is the Token-Passing, included in the family of deterministic access modes. In this method, a short message called Token is passed from a system to the other along the transmission mean. If the Token is marked free, the system receiving it is enabled to transmit a message. In this case, the system marks the Token busy, hooks it to the message and transmits it with the last. The message travels along the ring from a system to the other, taking the busy Token with it. Each system receiving the message checks the destination address to see if it must process the message. In each case, the system transmits the message and the busy Token to the next system. When the message reaches again the system it comes from, this same system provides to: • remove the message • mark the token free • transmit the Token to the next system. The Token starts traveling through the ring until another one needs to transmit. With this method the receiving station has the possibility to transmit messages within a particular period of time, after which it transmits the Token to the next station. If this one has no message to transmit, it passes the Token immediately. The most famous networks standards implementing this access method are the TokenRing/IEEE802.5 and the FDDI (see chapt.9.4).

5.4 TOKEN-BUS (IEEE 802.4) The Token-Passing is used in access control methods employed by Token-Bus technologies, too. In the Token-Bus, the systems are physically connected through a structure that can be bus, star or tree (the TokenRing allows only the ring structure). The systems use a logical ring structure for control of the Token passage. The logic ring is generally carried out with a system with addresses in decreasing order. From a logical point of view, the systems are connected to the transmission mean in a linear sequence, type A, B, C, D, etc. The Token can be passed according to the indicated sequence, starting from A to B, from B to C, from C to D and from D to A: a logical ring is formed in this way. The most famous networks implementing this access method are the Arcnet and other networks conceived for applications of industrial automation, based on the standard IEEE802.4.

- 48 -

5. ACCESSES CONTROL TECHNIQUES

5.5 QUESTIONS

Q1

What is the CSMA/CD code? 1 2 3 4

Q2

When two stations are simultaneously transmitting on a network there is a collision. Which of the following action is developed next? 1 2 3 4

Q3

the station or the stations which must receive the packet the station geographical position the part number of the network card the number chosen by the user to identify the station in the network

What is the token? 1 2 3 4

Q5

the station emitting the stronger signal will force the other to retransmit the packet the user provided with more privileges will have the line first the stations will generate a casual pause after they retry transmission the stations will be excluded by the networks after three attempts

What does identify the station address ? 1 2 3 4

Q4

single access data carrier detection multiple access data carrier detection, with collision identification data carrier removal when the collisions are multiple collisions removal via carrier removal.

a packet which can determine the number of stations present on the network a packet repeated from machine to machine to regulate the use of the network a packet containing a message addressed to all users a packet containing data, repeated by all stations to the destination

Which topology uses the token system most? 1 2 3 4

Star Bus Ring Braid

- 49 -

6. INTERNATIONAL STANDARDS

6. INTERNATIONAL STANDARDS 6.1 NETWORK ARCHITECTURE A network architecture defines the protocols, the formats of the messages and the standard which the machines and the software packets must conform to in order to reach specific purposes. As the network architectures are important, many organizations are interested in developing the standards. The different network architectures developed show a set of common objectives, which can be summed up in the following aspects: • connectivity • modularity • easy implementation • easy use • reliability • easy to modify. 6.2 ISO/OSI MODEL The International Standards Organization (ISO) defined a reference model for the Open System Interconnection (OSI). The OSI model provides a base to coordinate the development of standards concerning the flexible interconnection of systems employing data communication services. The OSI model divides the different possible phases of the information transmission into 7 levels. Each level develops a precise purpose, and the hierarchy goes from the higher to the lower level. The 7 levels of the OSI model are (fig.6.1): •

•

Level 7 – Application: this level provides those functions all application processes need for processing and information exchange. It is the only level that can be completely seen by the user, for which all the other levels are completely transparent. Files, mass memories, particular support functions are seen by the user as locally resident in the work station memory, although physically they can be found in other points of the network Level 6 - Introduction: it provides those services enabling the applications to interpret the exchanged data independently from the formats, the coding, the control characters and other manipulations carried out inside each system. So, if activated, it transforms the received information and show them to the user in a comprehensible way.

- 50 -

6. INTERNATIONAL STANDARDS

Fig.6.1 Levels of the OSI model

•

•

•

•

•

Level 5 - Session: it provides the services enabling the data exchange between application levels, establishing a logic connection without interruptions and errors. When the application level asks to fix a session and the destination process lies on a remote entity (typically a second processor}, the Session level accesses the Transport level, which fixes the necessary resources to carry out the connection Level 4 - Transport: its purpose is to activate and control a reliable data transport service between two systems. As it is responsible for the connection creation, it decides, e.g., how many channels must be used to ensure the required data flow (throughput}, or which is the less expensive and more reliable path among those possible. It provides control functions on the integrity of the received data, and controls the error data recovery Level 3 - Network: it provides the means to create, keep and abate the network connections required by the Transport level. Besides, it provides the procedural and functional means for routing and exchange of the Data Units Level 2 - Connection: it provides the functional and procedural means to transfer the Data Units among Network entities, and provides the correction and recovery of the data received wrong due to malfunctions of the physical level. Examples of Line Protocols are the BSC, HDLC, SDLC Level 1 – Physical: it provides the mechanical, electrical and functional means to activate, keep and disable the physical connections. Its purpose is to carry out the physical transfer of the binary data exchanged by the line level quantities.

- 51 -

6. INTERNATIONAL STANDARDS

6.1.2 Construction of the OSI frame Refer to fig.6.2. When the user A must send information to the user B, the data cross different levels of the model. Each level inserts a specific header containing service information for the homologous level, and generates a specific Data Unit of the same level (Protocol Data Unit – PDU). The conversation between homologous levels is defined by specific protocols for each level. Once the header is inserted, the level passes the PDU to the lower level and so on to the physical level, which purpose is to send the bits to the remote system. Once at destination, the frame will cross the different levels, and leave the part addressed to them.

Fig.6.2 Construction of the OSI frame

- 52 -

6. INTERNATIONAL STANDARDS

6.3 LAN IEEE MODEL The IEEE decided to standardize the networks when the first Ethernet and Token-Ring LAN appeared. Different Committees have been established, collected in the project IEEE 802. These Committees are: • • • • • • • • • • • • • •

802.1 802.2 802.3 802.4 802.5 802.6 802.3u 802.7 802.8 802.9 802.10 802.11 802.12 802.14

Overview, Architecture, Bridging and Management Logical Link Control CSMA/CD (Carrier Sense Multiple Access - Collision Detection) Token Bus Token Ring Metropolitan Area Networks - DQDB (Distributed Queue, Dual Bus) 100BaseT Broadband Technical Advisory Group Fiber-Optic Technical Advisory Group Integrated Data and Voice Networks Network Security Wireless Network 100VG Any LAN Cable-TV Based Broadband Communication Network.

The project IEEE 802 was developed in line with the OSI model, and describes the standards specific for the LAN related to the first 2 levels (fig.6.3). For the LAN applications, the Level 2 of the OSI (Data Link) is divided into 2 sublevels (fig.6.4): • LLC (Logical Link Control), enabling the access to the upper level services (Level 3, Network) independently from the access types of the same network, or the used physical support. The sub-level LLC is defined by the standard IEEE 802.2 • MAC (Media Access Control), controlling the access to the transmission mean. The sub-level LLC is common to all LANs, while the MAC is peculiar of each LAN, as well as the physical level it is strictly associated to. The sub-level LLC is the unified interface toward the Network level and is described in the proper standard IEEE 802.2, while the different MAC are described in the specific standards of each LAN (e.g. the MAC CSMA/CD is described by the standard IEEE 802.3).

- 53 -

6. INTERNATIONAL STANDARDS

7 6 5 4 3 2 1

Application Introduction Session Transport Network Data Link

IEEE 802

Physical

Fig.6.3 OSI and LAN (IEEE 802)

Level 3 Network

802.2

Level 2 Data Link

LLC

Logical Link Control

MAC Level 1 Physical

802.3

802.5

CSMA/CD

TOKEN RING

802.11

FDDI FDDI

Wireless

Fig.6.4 LAN and IEEE802

- 54 -

802.12 802.3u AnyLAN 100Base-T

6. INTERNATIONAL STANDARDS

6.3.1 Level 2 – LLC sub-level (Logical Link Control) The sub-level LLC, defined by the standard IEEE 802.2, sets as interface between Level 3 (Network) and the sub-level MAC (fig.6.5). The protocol used for the LLC sublevel belongs to the HDLC family, of which it is considered a variation for the LAN. The structure for the LLC Data Units (PDU, Protocol Data Unit) is similar to the one of the HDLC frame. It includes the following ranges (fig.6.6): • Destination Service Access Point (DSAP): 1 octect, it indicates the protocol of level 3 to which you pass the packet • Source Service Access Point (SSAP): 1 octect, it indicates the protocol of level 3 from which the packet comes • Control: 1 octect per Unnumbered frames, 2 octects per Information and Supervisory frames • Information: N octects, it contains the PDU of level 3

Fig.6.5 Relation between LLC-PDU and other levels

DSAP 1 octect

SSAP 1 octect

Control 1 or 2 octects

fig.6.6 Structure of the LLC-PDU

- 55 -

Information N octects

6. INTERNATIONAL STANDARDS

LLC addresses and the SNAPPDU

The LLC purpose is to provide a standard support to the coexistence of more upper level protocols (e.g. TCP/IP, X.25, DECnet, or others) on the same LAN, too. The addresses (of the services) contained in the DSAP and SSAP ranges are used to distinguish the protocol of Level 3 which the LLC frame is supported on (fig.6.7). As the 8 bits of the address field are sufficient to identify all protocols of Level 3, the ISO has given an official coding only to those protocols approved by a standardization body. E.g, the value FEH indicates the protocol ISO 8473 (Internet Protocol connectionless of OSI). When the DSAP and SSAP range of the LLC-PDU takes a value AAH there is a particular LLC frame, called SNAP (Sub Network Access Protocol) (fig.6.8). The SNAP frames are used to contain the PDU of Level 3 generated by proprietary protocols, i.e. not recognized by the ISO. In this case, the Control range indicates an HDLC frame of Unnumbered kind (the first two bits are 11), and is followed by a Protocol Identifier range composed by two parts: • The first 3 bytes contain the OUI (Organization Unique Identifier) of the organization that has proposed the protocol. E.g.: 00-AA00=Intel; 08-00-07=Apple) • the second 2 bytes identify the protocol inside the organization. If the first 3 bytes (OUI) are zero, the coding used for this range is the one defined by the Protocol Type of Ethernet 2.0. Example: − Protocol Identifier = 08 00 2B (OUI) 80 3C (Protocol Type) indicates a frame used by Digital (as 08002B is the OUI of Digital) for a proprietary protocol (indicated by 3C) − Protocol Identifier = 00 00 00 (OUI) 08 00 (Protocol Type) indicates a frame with Ethernet coding (OUI=000000) containing TCP/IP data, being 0800 the coding for IP.

Fig.6.7 Use of the DSAP and SSAP addresses

- 56 -

6. INTERNATIONAL STANDARDS

DSAP AAH

SSAP

Control

AAH

Protocol Identifier

03H

08 00 2BH OUI

DSAP AAH

SSAP

Control

AAH

80 3CH

L3-PDU

Protocol Type

Protocol Identifier

03H

Information

00 00 00H OUI

08 00H

Information L3-PDU

Protocol Type

Fig.6.8 Examples of SNAP frames

The LLC services

The sub-level LLC offers the following services at Network level: • Unacknowledged Connectionless Service (LLC Type 1): the data transfer does not set up virtual logic circuits between source and destination before carrying out the transmission, and needs no confirmation of the sent packets • Connection Oriented Service (LLC Type 2): the data transfer sets up virtual logic circuits between source and destination before carrying out the transmission, and needs confirmation of the sent packets • Semireliable Service (LLC Type 3): the data transfer does not set up virtual logic circuits between source and destination before carrying out the transmission, but needs confirmation of the sent packets

6.3.2 Level 2 – MAC sub-level (Media Access Control) A LAN is composed by more equipment disputing the access to a single transmission mean. The sub-level MAC (Medium Access Control) is specific of each LAN and solves the problem of the transmission mean sharing. There are different kinds of MAC, based on different principles such as the dispute, the token, the reservation and the round-robin. The MAC is necessary as at level 2 OSI (Connection) the LAN operates in broadcast, i.e. each system receives all frames sent by the other systems. To transmit in broadcast, i.e. make a single transmission channel be shared by all systems, implies the solution of two problems: •

•

in transmission: checking that the channel is free before transmitting, and solving any conflict between more systems needing to use the channel simultaneously in reception: determining which systems the message is actually addressed to, and which system has generated it.

The solution of the first problem is given by the different MAC algorithms, which must be distributed on different systems and must not need a master system. - 57 -

6. INTERNATIONAL STANDARDS

The solution of the second problem implies the presence of MAC levels addresses (so in the MAC-PDU) transforming broadcast transmissions into: • point-to-point transmissions (Unicast): if the destination address indicates a single system • point-to-group transmissions (Multicast): if the destination address indicates a group of systems • actually broadcast transmissions: if the destination address indicates all systems. The MAC-PDU structure is peculiar of each MAC, and, so, depends on the network. Some ranges, represented in fig.6.9, are present in all MAC-PDU, though. In particular, the generic MAC-PDU includes: • Destination Address: 6 octects • Source Address: 6 octects • Information: it contains the LLC-PDU • FCS (Frame Check Sequence): it contains the CRC value (Cyclic Redundancy Check) for control of the reception errors. Information Destination Address

Source Address

LLC-PDU

FCS

Fig.6.9 Structure of the generic MAC-PDU

The MAC addresses

The main function of the MAC address is to univocally identify the system on the LAN. The MAC addresses are univocal at world level, and are written permanently in a ROM of the NIC card by the manufacturer of the same card. They are composed of 2 parts of 3 bytes each (fig.6.10): • 3 most significant bytes: lot of addresses assigned by the IEEE to the card manufacturer (OUI, Organization Unique Identifier) • 3 less significant bytes, progressive production numbering, given by the card manufacturer (Vendor). E.g., a card with MAC 00-AA-00-2C-01-70H address is a card produced by the Intel, as 00-AA-00H is the OUI of this firm. 3 bytes

3 bytes

OUI

Vendor Assigned

Fig.6.10 Structure of the MAC addresses

The first 2 bits of the transmitted MAC addresses have a particular - 58 -

6. INTERNATIONAL STANDARDS

importance. Using a proper coding of the Destination Address, these 2 bits identify the kind of transmission carried out: • Unicast: toward a single system • Multicast: toward a group of systems • Broadcast: toward all systems When a card receives a packet, it does not pass it automatically to the upper level (LLC), but it carries out a set of controls: • it checks if the packet is integral (i.e., if the FCS is correct) • it analyzes the Destination Address: − if it is Broadcast, the packet is always passed to the LLC − if it is Multicast, the packet is passed to the LLC only if the card belongs to the addressed group, i.e. if the reception of that Multicast has been enabled by the software of upper level − if it is Unicast, the packet is passed to the LLC only if the Destination Address is equal to the card hardware address.

6.3.3 Level 1 - Physical The physical level provides the services to those users resident on the MAC sub-level. Figure 6.11 shows some of the functions included in the OSI model for the Level 1. The PDU exchanged with the physical level are signals representing the single bits composing a MAC frame. This level takes shape in the physical transmission of the signals on the transmission mean. It defines the procedures for carrying out the connections to the transmission mean, and for the signals transmission and reception. The specifications of the physical level include the description of the kind of cabling, the outlets and the connectors to be used, the characteristics of the signals to be exchanged.

- 59 -

6. INTERNATIONAL STANDARDS

6.4 REFERENCE MODELS As mentioned in the last chapter, the IEEE has divided the networks standardization into Committees. Hereafter follows a short description of the most used IEEE standards, such as the IEEE802, 802.3 and 802.5 (these standards will be treated exhaustively in the next chapters). Other non IEEE standards will be described, such as the FDDI and the ATM. 6.4.1 IEEE 802 The Project IEEE 802 has defined a flexible architecture oriented to the standardization of the LAN technologies. The approach maintained by the IEEE in the development of its network architecture fits the OSI model. As already seen, the Project 802, though, is focused only on the first two lower layers of the OSI model, the physical level and the Link level. The LAN architecture of the Project 802 has been accepted by the ISO and ANSI, too. 6.4.2 IEEE 802.3 IEEE 802.3 is the evolution of the Ethernet network, suggested before by Digital, Intel and Xerox (DIX). It uses a MAC type CSMA/CD (Carrier Sense Multiple Access/Collision Detection), in which the arbitrage of the transmission channel is regulated through a dispute mechanism. IEEE 802.3 includes a bus logic topology, with bus, star or tree cabling. The transmission speed is 10 or 100 Mb/s and the maximum throughput is about 4 or 40 Mb/s. 6.4.3 IEEE 802.5 IEEE 802.5 is the evolution of the Token Ring local area network suggested by IBM in alternative to Ethernet. The standard includes a ring topology, with star or double ring cabling. The arbitrage of the transmission channel is regulated by the token and, so, the protocol is deterministic, with stand-by time further restricted. The transmission speed is 4 or 16 Mb/s and the maximum throughput is 3 or 12 Mb/s. 6.4.4 FDDI (Fiber Distributed Data Interface) The standard ASC X3T9.5/ISO 9314 FDDI (Fiber Distributed Data Interface) defines a network with ring topology operating at the speed of 100Mb/s and using an access method type "token-passing" similar to what applied to the Token Ring LAN. Such standard specifies the use of full-duplex point-to-point connections in optical fiber technology to interconnect the stations, although implementations have been developed based on copper cable of not-shielded/shielded kind in "category 5". The variation of the FDDI for copper cable is often called CDDI (Copper Data Distributed Interface). - 60 -

6. INTERNATIONAL STANDARDS

The station interconnection to the FDDI ring can be carried out mainly in two modes (fig.6.11). • a first mode establishes that the station connects to the main ring with a single pair of optical fibers, of which one carries the signal in reception and the other in transmission. A network station using this kind of connection will be identified as station type SAS (Single Attachment Station) • a second mode establishes that the station connects to the main ring with two pairs of optical fibers, of which a pair is normally used to transport the signal in transmission/reception, and the other is available for back-up functions in case of interruption on normally used optical fibers. A network station using this kind of connection will be identified as station type DAS (Dual Attachment Station). This, in presence only of DAS stations on the main ring, takes to a physical network architecture which can recover any possible interruption, exploiting the inner path of the ring (normally not used).

Fig.6.11 FDDI ring

- 61 -

6. INTERNATIONAL STANDARDS

6.4.5 ATM (Asynchronous Transfer Mode) The set of ATM Forum standards defines a network: • with free topology (braid, star, tree) • operating at speeds ranging between 25 Mb/s and 2.5 Gb/s • using a method for sharing the transmission sources with definition of the traffic priority, such to guarantee well defined transmission capacity to the different users. Conceptually, the ATM technology can be considered an evolution of the packet switched networks, and it has been designed to meet the needs of modern high speed networks, heterogeneous in the structure as well in the supported traffic. The ATM network consists in a set of switching nodes (or switches) and in a set of terminal nodes (fig.6.12). The interface between a node and a terminal is called User-to-Network Interface (UNI), while the one between node and node is called Network-to-Network Interface (NNI).

Fig.6.12 Structure of ATM network

- 62 -

6. INTERNATIONAL STANDARDS

In ATM, an extremely fast packet switched model is implemented, totally carried out via hardware, in which the data are carried in units of fixed length called cells. Each cell is 53 bytes-long (fig.6.13) and uses only 5 bytes as Header, guaranteeing 48 bytes for the transport of the information flow (Payload). Two kinds of cells are defined with slightly different formats (fig.6.14): • UNI cell (User-to-Network Interface), used to transmit data between a user device and an ATM switching node • NNI cell (Network-to-Network Interface), used to transmit data between ATM switching nodes. ← Header ←

5 bytes

→

53 bytes Payload →

←

→

48 bytes

Fig.6.13 Structure of ATM cell

Header

GFC

VPI

VPI

VCI

VPI

VCI

- 5 bytes -

VCI VCI

VCI

Information

VPI

PTI

CLP

VCI

PTI

HEC

HEC

Information

Information

cella UNI

cella NNI

- 48 bytes -

GFC: GENERIC FLOW CONTROL VCI: VIRTUAL CHANNEL IDENTIFIER CLP: CELL LOSS PRIORITY

VPI: VIRTUAL PATH IDENTIFIER PTI: PAYLOAD TYPE INDICATOR HEC: HEADER ERROR CONTROL

Fig.6.14 Structure of UNI and NNI cells

- 63 -

CLP

6. INTERNATIONAL STANDARDS

The meaning of the Header ranges is the following: • GFC (Generic Flow Control): it is 4-bit long in the UNI cells, absent in the NNI cells. Its functions have not been defined, yet, but it should be used for a common control (generic) of the cells traffic for all kind of connection (data, audio, video, multimedia, …). Waiting for a definition, the 4 bits of the GFC range are set to 0000 • VPI (Virtual Path Identifier): it is 8-bit long in the UNI cells and 12bit long on the NNI cells. Together with the VCI, it identifies the virtual connection. The VPI constitutes a part of the identifier, and identifies a group of virtual channels (virtual path) • VCI (Virtual Channel Identifier): it is 16-bit long. Together with the VPI, it identifies the virtual connection. The VCI constitutes a part of the identifier and detects the virtual channels between two stations, inside a virtual path • PTI (Payload Type Indicator): it is 3-bit long. It indicates if the next Information range actually carries user data or service information for the network control • CLP (Congestion Loss Priority): it is 1-bit long. It indicates if the cell can be rejected by a switching node in case of congestion • HEC (Header Error Control): it is 8-bit long. It is the CRC (Cyclic Redundancy Check) calculated on the single Header. It enables the correction of the single errors, and the detection of double errors. The communication sessions in the network are activated defining the virtual circuits between different users; each pair of users in communication has a dedicated virtual circuit. In case of SVCs (Switched Virtual Circuits), the virtual connection will be active only for the time necessary to the information exchange between two users, and will be abated at the end of communication. In case of PVCs (Permanent Virtual Circuits), the connection between two users will be activated continuously, independently from the fact there is information exchange between them. To guarantee the different users/application the necessary band-width, priorities have been defined to be assigned to the different kinds of traffic supported by ATM. These priorities are called Quality of Service (QoS). At the base of the QoS concept, there is the fact that the applications are not all equal in terms of data network exploitation. E.g., in a videoconference service, a delay introduced in the data flow crossing the network causes voice asynchrony, or interruption of speech and image progress "by fits". On the contrary, for an application of "Office-Automation" taking the data from a network server system, a more or less contained delay in data loading does not affect its final efficiency. To identify and, so, to guarantee the different applications routed on the ATM network the necessary priority/usage band, some service classes have been defined to be assigned to the virtual circuits: • CBR: Constant Bit Rate • VBR-RT: Variable Bit Rate - Real Time - 64 -

6. INTERNATIONAL STANDARDS

• • •

VBR-NRT: Variable Bit Rate - Non-Real Time ABR: Available Bit Rate UBR: Unspecified Bit Rate.

To integrate such technology into local area networks (LAN) existing today, and to carry out local area networks in mixed technology, a standard has been defined called LAN Emulation (LANE). This makes the use of the ATM technology practically immediate to carry out backbone connections (backbone) between Ethernet networks, and ensures the inter-operability between Ethernet and ATM users. The standard LANE uses the encapsulation technique of the network card addresses at MAC level, as the last can support most network protocols of higher level. To implement such emulation, some specialized services have been defined: • LEC: LAN Emulation Client • LES: LAN Emulation Server • BUS: Broadcast and Unknown Server • LECS: LAN Emulation Configuration Server As concerns the transmission capacity of the communication interfaces of an ATM node, the last are many and enable high-speed data flows. The speeds available today for ATM products are: • OC-1 SONET: transmission at 51 Mb/s on optical fibers • OC-3 SONET: transmission at 155 Mbps on optical fibers • OC-12 SONET: transmission at 622 Mb/s on optical fibers • Carriers T1/T3: 1.544/44.736 Mb/s for transmission on conventional numerical flow in Extra-European countries • Carriers E1/E3: 2.048/34.368 Mb/s for transmission on conventional data flow in the European countries. With ATM technology, networks can be carried out having many of the characteristics of the local area networks of today, although offering a great technology for the service carriers, for the capacity to manage huge quantities of traffic with fixed delay times and service quality. The ATM technology can be used to create MAN/WAN networks. Summing up, the technique of "fast packet switching with fixed cell length" guarantees each system connected to the network to transmit a message within a predictable period of time. These performances are guaranteed in order to use the network with intensive applications or with applications which, by their nature, need to transmit flows of synchronized messages (multimedia applications, videoconference, etc.).

- 65 -

6. INTERNATIONAL STANDARDS

6.5 QUESTIONS

Q1

How many levels or layers is the OSI model composed of? 1 2 3 4

Q2

What does physical level indicate? 1 2 3 4

Q3

Media Access Control it controls the number of users effectively connected to the network the sub-level of the physical level of the OSI model it controls the sharing of the transmission mean on the network

Which is the main function of the sub-level LLC? 1 2 3 4

Q5

the functions developed by the electrical circuits (interface and cabling to the system) the protocol necessary to interconnect a network the applications used by the user to access the network the synchronization method of the network terminals

What is the sub-level MAC? 1 2 3 4

Q4

Two Four Seven Nine7

to provide a unified standard interface with the Network level to match the types of packets so to make them detectable from different systems to carry out a statistic on the packets transmitted by the users to interact with the physical level to control the network access

Which kind of network is identified by IEEE 802.3? 1 2 3 4

Token-Ring Ethernet Optical fiber Wireless

- 66 -

6. INTERNATIONAL STANDARDS

Q6

What topologies and speeds are included in the standard IEEE 802.5? 1 2 3 4

Q7

What does the ATM standard set define? 1 2 3 4

Q8

A network with free asynchronous topology with 16-Mb/s speed A network with synchronous braid with speed up to 2 Gb/s. A 2-Mb/s half-duplex infrared rays network A network with free topology with speeds ranging from 25 Mb/s to 2.5 Gb/s.

What does the SONET OC-3 specification identify ? 1 2 3 4

Q9

Star with 100-Mb/s speed Ring with 4/16 Mb/s speed Tree 10 Mb/s speed Bus with 100 Mbps speed

51-Mb/s optical fiber transmission on coaxial cable 155-Mb/s transmission on optical fiber Transmission for Extra-European countries in standard T1 or T3. A 2.048-Mb/s European carrier called E1/E3

Which kinds of network are good for ATM technology? 1 2 3 4

Small local area networks Geographical networks Satellite networks For local area as well as geographical networks

- 67 -

7. IEEE 802.3 - ETHERNET

7. IEEE 802.3 - ETHERNET 7.1 The ORIGINS of the ETHERNET In the first 70s, three industries (Digital, Intel and Xerox) created the pull of firms DIX for development of a local area network. At the end of the 70s, a first version of Ethernet was created, the version 1.0 operating at 10 Mb/s. In 1982, DIX issued the specifications of the Ethernet version 2.0, which was going to become the actual local area network. In parallel to the pool DIX, IEEE started the development of the standard 802.3. This is based on Ethernet, but differs for some logical characteristics referred to the Link level and for some electronic ones referred to the physical level. The reduced costs of the equipment and the great ease to design and carry out LAN of small/average dimensions have been the key of the success of Ethernet, and although all equipment on the market fit the specifications 802.3, they are often identified with the original name Ethernet. 7.2 The COMPONENTS Hereafter follow the description of the main components of a network 802.3, and, according to the needs, the details of the main differences between Ethernet and 802.3. 7.2.1 Transceiver 802.3 The Transceiver or MAU (Medium Attachment Unit) is the element enabling the transmission/reception of the packets between the interface (Controller) and the transmission mean, and changes according to the same transmission mean (fig.7.1). The interface is connected to the transceiver via a transceiver cable or drop-cable. In the cards 802.3, the transceiver can be integrated inside the Controller. The Transceiver mainly consists of: • receiving station: it transmits the data received by the transmission mean to the interface • collisions control section: it transmits a collision signal to the interface in case one has been detected on the transmission mean • transmission section: it receives the data from the interface and transmits them to the transmission mean • power supply section: it receives the power supply from the interface and generates the power supply of all electronic circuits inside the transceiver. In case of Ethernet Version 2.0, the only transmission mean allowed to connect the stations is the yellow coaxial cable (Thick cable). The first - 68 -

7. IEEE 802.3 - ETHERNET

transceiver has been developed to support the Thick cable as transmission mean, to which it is connected via a mechanical coupling system called "tap", perforating the cable by means of a point and touching the central conductor: such connection is known also as "vampire" connection. However, there is a variation where an "N" connector is used, with which the transceiver can be connected to a Thick cable provided with the same connector.

fig. 7.1 Connections between interface (Controller) and Transceiver

fig. 7.2 Connection between Transceiver and Thick cable

- 69 -

7. IEEE 802.3 - ETHERNET

7.2.2 Interface 802.3 The Interface 802.3 (or Controller) is the interface section between the inner bus of the network system and the transceiver. The interface develops the following functions (fig.7.3): • data encapsulation and de-capsulation • control of the connection to the transmission mean • bit Manchester coding and decoding.

fig. 7.3 Functions of the Ethernet interface – IEEE802.3

7.2.3 Drop cable The Drop cable, called also transceiver or AUI cable (Attachment Unit Interface) interconnects the transceiver to an Ethernet interface or to another repeater. It is a shielded cable with 15-pole connector and balanced signal transmission (see fig.7.1). 7.2.4 Repeater 802.3 The Repeater is used to extend the transmission mean length and carry out star or tree topologies. The active element interconnecting two coaxial cables or others transmission means is defined Repeater. In case of Thick cable, the Repeater needs two transceivers with AUI cable to connect the two segments. The main functions of an Ethernet repeater are: • to receive the strings of bits received on a segment and to transmits them to other segments by regenerating the signal • to decode (Manchester) the strings of received bits on a port, and recode them before re-transmission to the other ports - 70 -

7. IEEE 802.3 - ETHERNET

• •

• •

to control the collisions: if a collision is detected on a segment, the Repeater re-transmits it to the other ports to protect the segments from jabber errors (too long packets): when it detects a string of bits being transmitted for a period longer than 5 ms, it interrupts the transmission and resets it after a period of time ranging between 9.6 and 11.6 ms it can optionally insulate (partition) a port for a fixed period of time, when more than 30 consecutive collisions occur on the last it can house the integrated transceiver.

7.3 ACCESS METHOD The Ethernet and 802.3 networks are created with a bus topology based on coaxial cable, with transmission speed of 10 Mb/s, and involve the level 1 and the sub-level MAC of level 2 on the OSI battery (fig.7.4). Unlike the network IEEE802.3, where the Data Link level is divided into sub-levels MAC and LLC (802.2), in the Ethernet network there isn’t the LLC, and all functions of the Data Link level are developed by the MAC level. Level 3: Network

802.2

Ethernet

Level 1 Physical

LLC

Logical Link Control

Level 2 Data Link

802.3

802.5

CSMA/CD

TOKEN RING

FDDI FDDI

802.11 Wireless

802.12 802.3u

MAC

AnyLAN 100Base-T

Fig.7.4 Relations between the OSI model, Ethernet and IEEE802

The method used to split the use of the transmission mean between the network stations is the CSMA/CD (Carrier Sense Multiple Access / Collision Detection), identical in Ethernet and in 802.3. Is has been designed to use the coaxial cable as transmission mean, but it hasn’t been changed even after other transmission means - such as the optical fiber and the twisted pair - have been introduced.

- 71 -

7. IEEE 802.3 – ETHERNET

7.3.1 Control of the network access: CSMA/CD The main functions performed to access the network are described hereafter: • Carrier Sense: each station which must transmit listens to the bus and decides to transmit only if this is free (listen before talking) (fig.7.5-7.6, phase 1-2) • Multiple Access: it is possible that two stations decide simultaneously to transmit, as they can find the transmission mean free at the same time. The possibility this event should occur is increased by the fact that the propagation time of the signals on the cable is not null, and, so, a station can think that the mean is still free even when another has already started the transmission (fig.7.6, phase 1-2) • Collision Detection: if two transmissions should overlay, there is a "collision". To detect it, while transmitting a packet, each station listens to the signals on the transmission mean and compare them to one it generates (listen while talking) (fig.7.6, phase 3). After the collision detection, the following actions will be performed: • the transmitting station suspends the transmission and transmits the jamming sequence (transmission interference) composed by 32 bits for 802.3 and a number of bits between 32 and 48 for Ethernet: this sequence enables all stations to detect the occurred collision • the listening stations, detecting the fragment of collision consisting of the transmitted part of packet plus the jamming sequence, reject the received bits • the transmitting station repeats the transmission attempts after a pseudo-casual time, to a maximum of 16 times. This prevents the same stations to re-transmit at the same time after a collision. 7.3.2 Round Trip Delay A proper collision control made by the protocol CSMA/CD can be obtained if the transmitting station detects a collision. To achieve this result, the station must keep in transmission for a time long enough to enable any possible collision to propagate to it (it is proper to see that the collision can be detected only during the transmission, not after). Consider fig.7.7, where there are two stations (A and B) at the two ends of the network. In the worst case, there will be the following situation: • the station A finds the transmission mean free and transmits its frame (a) • the station B finds the transmission mean free, too, and transmits its frame an instant after the first bit of the frame coming from A reaches it (b) • B detects the collision (c) right after the beginning of the transmission • B suspends the transmission, and sets the jamming sequence (the

- 72 -

7. IEEE 802.3 – ETHERNET

•

transmitted part of packet plus the jamming sequence is called fragment of collision) in queue to the already emitted part of packet (d) in order to make A detect the collision, its transmission must last until the beginning of the transmission of B reaches A (e).

So, a transmission (i.e. a single frame) must last at least the time necessary for a bit to propagate from one end to the other of the network (from A to B), and vice versa (from B to A). This time is called Round Trip Delay. In case of Ethernet and IEEE 802.3 there is a: • transmission speed = 10Mb/s • minimum frame duration = 64 bit of preamble + 512 frame bits = 576 bits (see chapt.7.4), which corresponds to a minimum duration of 57.6 µs. The transmission duration of a packet is at least 57.6µs, and this is the maximum allowed Round Trip Delay. Half this time is the maximum propagation time of the signal from one end to the other of the network, that, at the propagation speed of 2·108 m/s, corresponds to a maximum extension of about 5 km. In practice, though, the attenuation introduced by the cables does not allow networks with such extension without the use of repeaters. 7.3.3 CSMA/CD and MAC sub-level The CSMA/CD access method and the sub-level MAC are responsible for the following operations: • packets transmission: during this phase the MAC accepts a packet with upper level and provides a serial string of bits to the physical level for their transmission • packet reception: the MAC receives a serial string of bits from the physical level and provides a packet to the upper level. If the packet has a different address from the one of the receiver, (or it is not a broadcast packet) it is rejected • transmission of a packet with busy channel in deferred mode: the MAC checks the errors in the packets comparing them to the value contained in the FCS range • the MAC guarantees there is minimum lapse of time called IGP (Inter Packet Gap) between two consecutive packets • the MAC interrupts the transmission when it detects a collision and transmits a "jamming" message after a collision • the MAC checks the minimum length of the packet (64 byte) • the preamble is generated in transmission and it is removed in reception.

- 73 -

7. IEEE 802.3 – ETHERNET

Fig.7.5 Transmission without collisions

Fig.7.6 Transmission with collisions

Fig.7.7 Round Trip Delay

- 74 -

7. IEEE 802.3 – ETHERNET

7.4 PACKET FORMAT at MAC LEVEL The Ethernet packet and the packet 802.3 slightly differ between them (fig.7.8). The packets (excluded the Preamble and SFD ranges) have a variable length between 64 and 1518 bytes (or octects), corresponding to 512 bit (minimum) and 12144 bit (maximum). Hereafter follow the description of the different ranges, and the main differences between Ethernet and 802.3: •

•

•

• • • •

•

•

Preamble: sequence of alternated “1” and “0”, 7-octect long in Ethernet as in IEEE 802.3. The Ethernet frame includes an additional byte, which is the equivalent of the SFD byte of the frame 802.3. The preamble is used to synchronize the receivers of the network stations Start-of-Frame Delimiter: a sequence of 8 bits with value "10101011" (in 802.3), indicating the beginning of a data frame. In Ethernet, the "Preamble" and the "Start delimiter" range are grouped and called simply "Preamble" Destination Address: 6 octects, it identifies the station or the stations which must receive the frame. The destination address can specify a single (unicast) as well as a group of addresses (multicast or broadcast) (see chapt.6.3.2) Source Address: 6 octects, it identifies the station originating the frame Length – IEEE 802.3: 2 octects, it indicates the number of octects contained in the Data field (LLC-PDU) Type - Ethernet: 2 octects, it indicates the kind of protocol of upper level generating the PDU contained in the Data range Date: it contains the information exchanged between the network stations. In IEEE 802.3, if the data are not sufficient to obtain a frame of at least 64 byte, filling octects (PAD) are inserted. Ethernet expects at least 46 data bytes Filling field (PAD) – IEEE 802.3: used only in IEEE 802.3, it contains a fixed number of filling bytes to ensure the frame a minimum length of 64 bytes Frame Check Sequence: 4 octects, it contains the value of the cyclic redundancy control (CRC) for error detection.

Note there isn’t a packet end signaler. The separation is carried out by inserting an Inter Packet Gap between two consecutive packets.

- 75 -

7. IEEE 802.3 – ETHERNET

Preamble 7 octects

SFD

Destination Address

Source Address

Length

1

6

6

2

Data

PAD

FCS

from 0 to 1500 from 0 to 46 4

IEEE 802.3

Preamble 7 octects

SFD

Destination Address

Source Address

1

6

6

Type 2

Data

FCS

from 46 to 1500

4

Ethernet

fig.7.8 Format of the frames IEEE 802.3 and Ethernet

- 76 -

7. IEEE 802.3 –ETHERNET

7.5 The PHYSICAL LEVEL We report hereafter the characteristics of the different transmission means supported by the specification IEEE 802.3 and Ethernet (only Thick cable, see following 10Base5). To define some important characteristic aspects of a network (speed, Band Base or Translated Band transmission technique, transmission mean or maximum distance), some codes are used nowadays which meaning is reported hereafter: 10BaseT transmission mean transmission technique speed

- 77 -

T = twisted pair 2 = (thin) cable for max. distance of 200 m 5 = (thick) cable for max. distance of 500 m F = optical fiber

7. IEEE 802.3 –ETHERNET

7.5.1 10Base5 The specifications of this standard concern the characteristics of the MAU (transceiver) and the transmission means for speeds of 10 Mb/s, based on a segment of 500 m at max. The transceiver 10Base5 can transmit and receive electrical signals along a 500m Thick coaxial segment, to which at maximum 100 Transceivers can be connected set at a minimum distance of 2.5m between them (fig.7.9). The segment 10Base5 consists of a 50-Ohm coaxial cable type RG8 (called also Thick or yellow cable). The connection to a station and a segment of cable occurs by means of the MAU mechanical coupling based on "vampire" or type "N" connector (fig.7.2). The path between any pair of stations can be implemented by using 5 500-m segment of cable and 4 Repeaters at max.

a = 2.5m minimum distance between 2 transceivers b = segment length = max 500m is a single trunk = max 491.4m if more trunks connected with joints

fig.7.9 Configuration of the segment 10Base5

- 78 -

7. IEEE 802.3 –ETHERNET

7.5.2 10Base2 The specifications of this standard concern the characteristics of the MAU (transceiver) and the transmission means for speeds of 10 Mb/s, based on a segment of 185 m (about 2 100-m units). The Transceiver 10Base2 (fig.7.10) can transmit and receive electrical signals along a 185-m Thin coaxial segment, to which up to 30 Transceivers can be connected set at a minimum distance of 0.5m between them (fig.7.11). The segment 10Base2 consists of a 50-Ohm coaxial cable type RG58 (called also Thin cable). The connection of a station to a segment of cable occurs by means of the Transceiver and a mechanical coupling system based on the "T" connector type BNC. Very often, the Transceiver is integrated into the network card. The path between any pair of stations can be implemented by using at max. 5 185-m segments of cable, and 4 Repeaters.

fig.7.10 Connections of the Transceiver 10Base2

a = 0.5m minimum distance between 2 transceivers b = segment length = max 185m

fig.7.11 Configuration of the segment 10Base2

- 79 -

7. IEEE 802.3 –ETHERNET

7.5.3 10BaseT The specifications of this standard concern the characteristics of the MAU (Transceiver) and the transmission means for 10-Mb/s speeds, based on a 100m Twisted Pair segment. The Transceiver 10Base2 can transmit and receive electrical signals along a 100m twisted segment between 2 stations in point-to-point mode. The peculiarity of this standard is the use of multi-port repeaters (Hubs) to connect more than 2 network stations, with star or tree topology (fig.7.12). Up to 4 Hubs can be connected in cascade. The maximum distance between stations and Hub is 100m, and more than 1000 stations cannot be connected. The segment 10BaseT consists of a 100-Ohm twisted cable type UTP, STP or FTP, which specifications are the same required by the structured cabling systems. The connection of a station to a segment of cable occurs by means of the MAU, integrated into the network card, and of a mechanical coupling system based on the 8-pole connector "RJ45" (fig.7.13). The path between any pair of stations can be implemented by using at max. 5 100-m segments of cable and 4 Repeaters (Hub).

fig.7.12 Network 10BaseT architecture

fig.7.13 Connector RJ45

- 80 -

7. IEEE 802.3 –ETHERNET

7.5.4 10BaseF (FB and FL) The specifications of this standard concern the characteristics of the MAUs (Transceivera) or the transmission means for 10-Mb/s speeds, based on a 2000-m segment optical fiber. The Transceiver 10BaseFB (Fiber Backbone) can transmit and receive electrical signals along a 2000-m optical fiber segment between 2 Repeaters in point-to-point mode (the connection between 2 stations or between station and Repeater is not allowed), on connections with backbone function. The segment 10BaseFB consists of an optical fiber cable 62.5/125µm headed on ST or SC connectors. The connection between Repeaters occurs through the MAU. The MAU 10BaseFL (Fiber Link, compatible but with better performances than the last FOIRL – Fiber Optic Inter Repeater Link) can transmit and receive electrical signals along a 2000-m optical fiber segment between a station and a Repeater or between 2 stations in pointto-point mode (fig.7.14). The connection between 2 Repeaters is not included. The segment 10BaseFL consists of an optical fiber cable 62.5/125µm addressed on ST or SC connectors. The connection between station and Repeater or between station and station occurs by means of the MAU.

fig.7.14 Connection between two 10BaseFL MAUs

- 81 -

7. IEEE 802.3 –ETHERNET

7.6 QUESTIONS

Q1

Which is the maximum length of the segment in the standard 10Base5 ? 1 2 3 4

Q2

Which is the maximum number of stations which can be connected to a 10Base2 segment ? 1 2 3 4

Q3

from 1 to 64 octects from 32 to 32768 octects from 64 to 1518 octects from 128 to 256 octects maximum of 64 octects

How long is the address field in the Ethernet frame ? 1 2 3 4 5

Q6

type of cable used in cabling speed supported by the network maximum number of users type of modulation

Which is the length of an Ethernet packet ? 1 2 3 4 5

Q5

5 100 30 20

What does the word Base present in the codes mean ? 1 2 3 4

Q4

500 m 100 m 50 m 25 m

5 octects 6 octects 7 octects 8 octects 6 bits

What is a jamming sequence ? 1 2 3 4

transmission of a particular form of packet by the station detecting the collision message transmitted by an out-of-use station reset sequence after a collision carrier over-modulation.

- 82 -

8. IEEE 802.3u – FAST ETHERNET

8. IEEE 802.3u – FAST ETHERNET 8.1 100BaseTX and 100Base T4 The network 100BaseT (or IEEE 802.3u) is currently the only LAN which can be defined "Ethernet at 100 Mb/s", as it keeps the classic algorithm CSMA/CD implemented on 10BaseT as it is, but operating at 100 Mb/s. The structure and the minimum length of the MAC frame have not been changed and, so, the Round Trip Delay has been reduced of a factor 10, determining the extension of the collision domain and so the maximum dimensions of the network. 100BaseT uses the existing interface of level MAC IEEE 802.3 and connects it, across a layer called MII (Medium Independent Interface), to a physical sub-levels family (transmission means) including: 100BaseT4, 100BaseTX and 100BaseFX. The pair TX/FX is based on the interfacing technology to the transmission mean already used in FDDI, and transmits with 4B5B coding (see chapt.4.6) at 125 Mb/s, on 2 UTP pairs of category 5 (variation TX, fig.8.1a) or on 2 multimode fibers (variation FX). 100BaseT4, instead, uses a new interfacing technology to the transmission mean, transmitting with 8B6T coding (see chapt.4.6) at 100 Mb/s on 3 UTP pairs of category 3 (variation T4, fig.8.1b): the fourth pair is used to control the collisions.

Fig.8.1 Use of the pairs in the variations TX and T4

- 83 -

8. IEEE 802.3u – FAST ETHERNET

Due to the increase of a factor 10 on the transmission speed, to the maintenance of the protocol CSMA/CD and to the format of the packets IEEE 802.3, the collision domain reduces of a factor 10 in respect to the Ethernet-802.3 network. So, the maximum allowed distance between the two stations reduces to 210m (fig.8.2). This, however, enables the cabling of 100BaseT around a hub with 100m of ray, and, consequently, 200m of diameter: 100BaseT, so, it is compatible to the standards for structured cabling. The objective of 100BaseT is to keep the compatibility to 802.3 at cards level using exactly the same packet format, and to have a very interesting economic positioning; in average, the 100BaseT products cost only the 50% more than the analogous 10BaseT ones. To support the economy of the standard, many devices nowadays offer the possibility to switch the speed of the port between 10 and 100 Mb/s as function of the connected device. The mechanisms regulating the port are called N-Way Auto-Negotiation (NWAN) and Auto Sensing (AS).

Fig.8.2 Structure of 100BaseT network

- 84 -

8. IEEE 802.3u – FAST ETHERNET

8.2 QUESTIONS

Q1

Which maximum speed is supported by a Fast Ethernet network ? 1 2 3 4

Q2

Which kind of cable has been approved for this kind of network? 1 2 3 4

Q3

Proprietary Token-Ring CSMA/CD Token-Passing

How many optical fibers does the variation 100BaseFX use? 1 2 3 4

Q6

100 m 200 m 500 m 12 Km

Which access method is it used by the 100BaseT networks? 1 2 3 4

Q5

Twisted pair and optical fiber Only optical fiber Coaxial cable Radio waves

Which is the maximum diameter of a 100BaseTX network? 1 2 3 4

Q4

52 Mb/s 2 Gb/s 16 Mb/s 100 Mb/s

One Two Four Eight

Which signal coding is used by the 100BaseT4 standard? 1 2 3 4

4B5B 8B6T 6B8T 9B2B - 85 -

9. IEEE 802.5 – TOKEN RING

9. IEEE 802.5 – TOKEN RING 9.1 The ORIGINS of the TOKEN-RING The Token-Ring was born in 1976 in the IBM laboratories, as local area network alternative to Ethernet. It has been created to operate on a star cabling carried out with STP cable. The first version has a speed of 4 Mb/s. In 1982, the IEEE created the Committee IEEE 802.5 which purpose was to draw up a Token Ring standard relatively to the physical level and the sub-level MAC of the Data Link level. The Committee introduced some changes and introduced a 16-Mb/s version, too, using passive concentrators and STP cables. In 1993, a proof was issued offering, in particular, the possibility to use UTP cables. 9.2 The ACCESS METHOD A Token Ring consists in a particular number of stations connected in series via a transmission mean and closed as ring (fig.9.1). The packets are transferred from a station to the other serially: each station repeats and regenerates the transmission toward the next station. As the stations must continuously repeat the packets of the other stations, the network is star cabled for more reliability. The connections between the star center and the stations are called lobes. When a station is off or faulty, the star center (concentrator, or MSAU, MultiStation Access Unit) excludes it from the network.

Fig.9.1 Logical diagram 0 0 of Token-Ring network

- 86 -

9. IEEE 802.5 – TOKEN RING

The MAC access method is the token. The token is a particular type of packet flowing across the ring indicating that the ring is free (fig.9.2). A station wanting to transmit must wait for the token, capture it, mark it busy and then transmit the message together with the token. The token continuously flows through the ring although the stations do not have data to transmit. It is generated at start by the station that has earned the role of network Active-Monitor and is repeated by all stations. A station that has captured the token can transmit one or more packets as function of their length and the Time Holding Token parameter (THT) indicating the maximum time a station can keep the token.

Fig.9.2 Use of the Token

9.3 TOKEN and FRAME FORMAT The token is composed by three bytes (fig.9.3): • • •

Starting Delimiter (SD) Access Control (AC) End Delimiter (ED).

These three ranges have analogous format and functions for the token and the frames. The meaning of the different ranges is the following (fig.9.4-9.5): • Start of Frame Sequence (SFS): it indicates the start of the frame and is composed by a Starting Delimiter byte and a second Access Control byte • End of Frame Sequence (EFS): it indicates the end of the frame and is composed by an Ending Delimiter byte and a second Frame Status byte • Starting Delimiter (SD): it identifies the starting of the token or the frame. On this purpose, it contains the bits identified as "J" and "K", violating the Differential Manchester coding (used in Token-Ring)

- 87 -

9. IEEE 802.5 – TOKEN RING

•

• •

Access Control (AC): it contains the access information to the ring. The "token bit” takes the value "0" in case of token and "1" in case of a frame Ending Delimiter (ED): it indicates the end of the packet. Frame Status (FS): it contains the Address Recognized and Frame Copied bits: the combinations of these 2 bits indicate: − not-existing or inactive station in the ring (A=0, C=0) − existing station, but the frame has not been copied (A=1, C=0) − the frame has been copied (A=1, C=1).

The actual packet starts after the SFS and can have a length between 21 and 17796 bytes: • Destination Address (DA): address of the frame destination • Source Address (SA): address of the frame source • Routing Information (RI): it contains the routing information for extended local area networks, according to the IBM proprietary routing protocol known as Source Route Bridging • Frame Control (FC): it defines the packet content differentiating it between MAC frame (for network control) or the LLC/PDU (information frame) • Frame Check Sequence (FCS): it contains the CRC frame. Fig.9.3 Format of the Token IEEE 802.5

Fig.9.4 Format of the frame IEEE 802.5

Fig.9.5 Format of the SD, AC, FC, ED, FS ranges

- 88 -

9. IEEE 802.5 – TOKEN RING

9.4 NETWORK ACCESS CONTROL Consider the example of fig. 9.6. The transmission of a frame from station “A” to station “C” is made as follows: •

• • • •

• •

•

•

•

the station “A” waits in order to receive the token and captures it (fig. 9.6a). The capture occurs by taking to "1" the "token bit" of the AC range. This modification transforms the token into a frame and, in particular, the already transmitted part becomes the SFS “A” inhibits the bit repetition circuit between input and output “A” transmits the FC, the DA, the SA and eventually RI for the Bridge Source Routing “A” transfers the data into the Info range (fig. 9.6b) if the station “A” has other frames to transmit but is has not overcome the THT, yet, it sets to 1 the intermediate bit of ED and starts the transmission of the next frame when “A” has transmitted the last frame, it sets the intermediate bit of ED to 0 if “A” ends the frame transmission before having started to receive backward from the input, it transmits the filling bits on the output until it can regenerate the new token when “A” receives the SA range of the transmitted frame on the input port, and detects it as its own, it removes the frame from the ring and starts to emit the token (fig.9.6c). If the frames transmission is ended, it immediately generates the new token, on the contrary it waits for the end of transmission all stations that do not have the token (in this case B, C and D) repeat the bits they receive to the next station. Each station observes the frames it receives to check if the destination address DA is equal to the its MAC address. This equality occurs only on the station the frame is addressed to (the station “C” in this example, fig.9.6d). Besides, it not only repeats the frame, but it receives and changes the address recognized bit and the copied bit properly, in the FS range at the end of reception of its frames, the station A restores the repetition of the bits received between the input and output ports.

- 89 -

9. IEEE 802.5 – TOKEN RING

a. Token capture

b. transmission

c. new Token generation

d. frame reception and copy

Fig.9.6 Frames transmission on Token-Ring

- 90 -

9. IEEE 802.5 – TOKEN RING

9.5 NETWORK CONTROL 9.5.1 Functional addresses As already seen, the addresses of the source and destination are contained in the SA and DA ranges, and each is 6-bytes long. There are some "multicast" MAC addresses (called functional addresses) used to develop the following 4 functions: • Active Monitor (C0 00 00 00 00 01): function developed by the station generating the token and fixing the clock for the other stations • Ring Parameter Server (C0 00 00 00 00 02): function for initialization of the parameters related to the active stations in the ring • Ring Error Monitor (C0 00 00 00 00 08): function which can collect errors from the stations, analyze them and generate statistics • Configuration Report Server (C0 00 00 00 00 10): function receiving the stations configuration information. 9.5.2 Active Monitor Beside the normal activity connected to data exchange, in a Token Ring network there is a heavy control activity. In a ring, there is a single station developing functions of Active Monitor for the network, and is designed for this function after a token claim process. All the other stations set in Stand-by Monitor state, ready to become the Active Monitor of the network in case of problems on the existing Active Monitor. The Active Monitor communicates every 7 seconds its presence to all network stations via an Active Monitor Presence (AMP) packet. If a station in Stand-by Monitor state does not detect an AMP packet transit within this time, the election process will be started which will take to the definition of a new Active Monitor. 9.5.3 Control activity Among the most important control activities, there are the Duplicate Address Test, Neighbor Notification, Ring Purge Beacon Process activities. The Duplicate Address Test is used by all stations each time they enter the network, to check that no station with the same address is active on the network. To do so, when entering the network the station sends a frame with its DA address: if the frame comes without alterations (A=0, C=0) it means that there is no address duplication. The Neighbor Notification process is controlled by the Active Monitor. This process enables each station to know the address of the station

- 91 -

9. IEEE 802.5 – TOKEN RING

connected to the input, i.e. the Nearest Active Upstream Neighbor (NAUN) or Upstream Neighbor Address (UNA). Even the Ring Purge process is controlled by the Active Monitor: when it detects there isn’t the token or there is a too long transmission on the network, it provides to send a ring purge frame, so that the stations reset the network interface and start to operate properly, again. In presence of a fault, though, the ring purge process can fail: in this case the Beacon Process activates. The fault insulation process is activated when the Active Monitor election process is activated.

9.6 The PHYSICAL LEVEL The Token-Ring physical level changes as function of the considered standard. The version ISO 8802.5 defines the use of the 150-Ohm STP cable and of passive concentrators called MSAU (MultiStation Access Unit). The 802.5 Q/Draft 3 enables the use of UTP cables of category 3, 4 and 5 whenever the speed is reduced to 4 Mb/s and of category 4 or 5 for the 16 Mb/s. 9.6.1 The cabling The path of an information of a Token-Ring network is ring, but the cabling between stations and concentrator (MSAU) is star. Lobo is the connection between concentrator and station, Ring-In and Ring-Out (or Backbone) the connections between concentrators. The concentrators are connected to counter-rotating double ring via primary ring and a backup one (fig.9.7a). In case of interruption of the primary ring, the backup ring keeps the connection between the MSAU (fig.9.7b). The physical connection appears on the station side with connector DB9 and on the concentrator side via connector hermaphrodite or IBM Data Connector (IDC). If the cabling should be carried out with UTP cable there are the Media Filter (or baloons) which purpose is to match the impedance from 150 to 100 Ohm.

- 92 -

9. IEEE 802.5 – TOKEN RING

a. connected primary and backup rings

b. cut off primary ring

Fig.9.7 Example of Token-Ring network

- 93 -

9. IEEE 802.5 – TOKEN RING

9.6.7 The concentrators (MSAU) The MSAU can have a number of ports between eight and twenty and can operate at 4 and 16 Mb/s. They are divided into three categories: • Passive: they don’t have automatic bypass mechanisms for the faults on the Ring-In and Ring-Out ports (backbone). They have insertion and by-pass detection mechanisms on the lobe ports • Active: they are equipped with amplification and re-timing circuits on each port and are used on UTP cables. They control the bypass on the Ring-In and Ring-Out ports • Partially active: as the last, but only on the backbone ports. When a station is active, it can ask the concentrator to be inserted into the ring. On this purpose, it generates a voltage between 3.5 and 7V, called insertion voltage. This voltage enables a relay inserted into the concentrator (one for each lobe) to switch from OFF (station excluded) to ON condition (station inserted). The station has an access circuit to the ring providing or removing the insertion voltage according to whether the station should be inserted into the ring or it should be in bypass for test purposes or due to faults.

- 94 -

9. IEEE 802.5 – TOKEN RING

9.7 QUESTIONS

Q1

Which are the transmission means good for the standard IEEE 802.5 ? 1 2 3 4

Q2

How many categories are the MSAU divided into ? 1 2 3 4

Q3

A primary and a secondary A primary and a backup secondary A primary A primary, a secondary and a backup one

How many circuits are the active MSAU equipped with ? 1 2 3 4

Q5

Two: passive and active Three: passive, active, partially active Three: passive, positive, negative Four: passive, active, alternative, positive

How many rings is the Token-Ring network composed of ? 1 2 3 4

Q4

UTP cable, optical fiber Only UTP Only optical fiber Coaxial cable, optical fiber

Amplification, signal re-timing and bypass circuits Bypass and signal amplification circuits Signal switching and re-timing circuits Re-timing and bypass

How long is the token ? 1 2 3 4

24 octects 3 octects 9 octects 64 octects

- 95 -

9. IEEE 802.5 – TOKEN RING

Q6

What is the active monitor ? 1 2 3 4

Q7

Which coding is it used in Token-Ring transmission? 1 2 3 4 5

Q8

An instrument to evaluate the network performances The station generating the token and synchronizing the remaining ones The station controlling the token once generated and synchronizing the remaining ones A MSAU with network supervision purposes

MLT-3 NRZI RZI Differential Manchester Manchester

What must a station do to start transmission? 1 2 3 4

Generate the token and tune the message octects Capture the free token and tune the message octects Wait for the next token to tune the message octects Wait to be appointed active monitor

- 96 -

10. EVOLUTION of LOCAL AREA NETWORKS

10. EVOLUTION of LOCAL AREA NETWORKS 10.1 SWITCHED NETWORKS The spreading of multimedia applications, that must transfer not only the data but also the flows coding voice and video services, requires an increment of the band (data flow in bit/s) available for the single work area. All examined LANs expected the existence of a single transmission mean with high speed and low error rate, which transmission or band capacity was shared by all systems connected. We have seen that the standards related to the structured cabling system took all LANs back to a star or tree topology. The star topology does not bring advantages in terms of LAN total transmission capacity if the concentrators behave as normal Repeaters or Hubs (Ethernet, IEEE 802.3) or simple star centers (Token-Ring). Changing the Hub concentrators with frames switches at the MAC sublevel (level Data Link), you reach a total transmission capacity higher than the one of the single connections between a system and the Hub. Such frames switches are called Switches (fig.10.1-10.2). A Switch has the capacity to transmit more packets simultaneously if sources and destinations are different, generating a number of temporary virtual connections dedicated to the union of two or more interconnected stations. 10.1 Ethernet "switching" The term "Ethernet switching" indicates an Ethernet network in which there are Switches instead of normal Hubs. From a logical point of view, the Switches are multiport bridges at all effects. When receiving a MAC frame, the Switch examines immediately the destination address, consults its routing tables to determine the destination port, and, if this is free, starts re-transmitting the frame while still receiving it: this mode is known as Cut-trough. When in this mode, the Switch does not recalculate the frame FCS, as the FCS is set at the end of the frame, and so, unlike the Bridge, it cannot prevent corrupted frames to be sent on the LAN. A second method consists in the Store-and-Forward technique, as for the Bridges. This technique can be used when the Switch operates in two separate local area networks with different standards, identical networks with different speeds, when the destination port is often busy, when the LAN is subject to many errors or when the packet is broadcast kind.

- 97 -

10. EVOLUTION of LOCAL AREA NETWORKS

Fig.10.1 From the concentrator to the Switch

Fig.10.2 The Switch operates at Data Link level

- 98 -

10. EVOLUTION of LOCAL AREA NETWORKS

10.2 VIRTUAL LOCAL AREA NETWORKS (VLAN) The virtual local area networks technology (Virtual LAN or VLAN) refers to the capacity offered by the Switches or the Router to configure more logical networks over a single physical local area network (fig.10.3). Each VLAN consists of a set of segments of local area network that can include a single station or group of stations. The stations belonging to a VLAN are logically interconnected to Data Link level, although they are physically connected to different transmission means. The main advantages that can be obtained from this assignment come from the traffic insulation of the different work groups at Data Link level. This is not only important for safety and data privacy reasons, but also because it allows keeping the multicast/broadcast traffic separated in the different virtual networks. The inter-operability between VLANs is guaranteed by an external unit, usually a Router. Many manufacturers suggest the possibility to create VLAN or users domains on their top Hubs/Switches, although this functionality is often limited to a single Hub/Switch. To make the VLAN concept really useful, it must be possible for a domain to include ports belonging to different Hubs/Switches (fig.10.4).

Fig.10.3 Virtual networks on the same Hub

Fig.10.4 Virtual networks on more interconnected Hubs

- 99 -

10. EVOLUTION of LOCAL AREA NETWORKS

10.3 QUESTIONS

Q1

Which is the function carried out by the switch? 1 2 3 4

Q2

Which is the parameter the switch is chosen with ? 1 2 3 4

Q3

3 4

3 4

The switch memorizes the packet to route it The switch receives a packet and routes it during the reception phase toward the destination port The switch memorizes a particular number of packets and reconstructs the routing tables Packets routing technique used only on the bridges

In which situation is it advisable to use a switch ? 1 2 3 4

Q6

The switch cuts the packets it is receiving by removing them The switch receives a packet and routes it during the reception phase toward the destination port The switch memorizes a particular number of packets to send them to the destination ports The switch takes the packet and transmits it to the destination port with shorter path

What is the store-and-forward technique? 1 2

Q5

Pass band Number of ports Supported voltage Switching speed

What is the cut-through mode? 1 2

Q4

To join all hubs of a network To join all PCs of a network by switching them manually To generate virtual connections between two high traffic stations To join two networks operating with different network operating systems

When the number of stations is over 100 When there is an intense traffic between stations When there is a large number of hubs To carry out a geographical network

What does the term VLAN mean? 1 2 3 4

A mixed local area network A virtual local area network A wireless local area network A geographical network. - 100 -

11. UPPER LEVEL PROTOCOLS

11. UPPER LEVEL PROTOCOLS 11.1 The TCP/IP PROTOCOL TCP/IP is a set of transmission and routing protocols widely used in the world of local area networks. The acronym 'Transmission Control Protocol/Internet Protocol' comes from the main protocols included in the set, developed as research project in 1969 by the Department of Defense of the United States of America. The TCP/IP protocols are based on the packet switched network, and enable single networks of interconnected systems to appear as a single network, called "internet", where all systems can freely exchange data between them as if directly connected. Today the TCP/IP is used in the whole world on a very large number of local area networks: first of all InterNet, which nowadays connects thousands of networks containing millions of systems including universities, national laboratories and commercial organizations. The network software implementing the main TCP/IP protocols is available on a large range of information systems: from the Mainframe up to the Personal Computers. The TCP/IP network architecture does not specify the physical levels and Data Link of the network, but it uses those normally available and fitting the standards. E.g., in the field of local area networks, it operates on Ethernet and IEEE 802.3, Token-Ring and IEEE 802.5, FDDI, ATM, etc.. In the field of geographical networks it operates on PPP, Slip, ATM, etc.

Fig.11.1 TCP/IP Architecture

- 101 -

11. UPPER LEVEL PROTOCOLS

IP addresses

Each system connected to an TCP/IP internet is identified by an univocal address in 32-bit format. The IP addresses, which must be univocal on the network, are expressed by writing the decimal values of each byte separated by a point. Examples of IP addresses are: 200.1.1.1 221.191.61.70 46.0.1.121 The starting bits of the address identify its typology and describe the way in which the 32 address bits are divided between network (network) and host range (system). The network range detects the single network on which the origin and destination system are located, while the host range identifies a specific system of that network.

Classes

The starting bits of the network range specify the address class they belong to. The most used classes are: • Class A: it is an address where the first bit is "0". An address in class A provides 7 bits to identify the physical network and 24 bits to identify the systems belonging to the network. The addresses of class A can be detected as the first address range is between 0 and 127 • Class B: it is an address where the first two bits are "10". An address in class B provides 14 bits to identify the physical network and 16 bits to identify the systems belonging to the network. The addresses of class A can be detected as the first address range is between 128 and 191 • Class C: it is an address where the first three bits are "110". An address in class C provides 21 bits to identify the physical network and 8 bits to identify the systems belonging to the network. The addresses of class A can be detected as the first address range is between 192 and 223 • Class D: it is an address where the first four bits are "1110". An address in class D is used to carry out a particular broadcast shape (multicast), in which an address identifies a set of systems. The addresses of class A can be detected as the first address range is between 224 and 239

Fig.11.2 Classes of the TCP/IP addresses

- 102 -

11. UPPER LEVEL PROTOCOLS

Netmask

The host part of an address of class A, B or C can be divided into two parts, subnet and host. The amplitude of the subnet and the host range is defined via a parameter called netmask. The netmask contains bits 1 in correspondence to the network and subnet ranges, and bit 0 in correspondence of the host range. E.g., a netmask 11111111 11111111 11111111 00000000 , commonly written as IP address 255.255.255.0, indicates that the host range coincides to the last address byte. The use of the subnets and the netmask makes the address detection and, consequently, the data routing quicker.

11.2 The IPX/SPX PROTOCOL The protocols IPX/SPX of Novell (NetWare, fig.11.3) are based on Xerox protocols and have a levels structure. The protocol IPX (Internetwork Packet Exchange) is the level Network protocol and provides a service type "datagram" without connection, not reliable to control the information exchange between the different systems. Its purpose is to deliver a packet to the destination but does not need confirmation that the packet has been received. IPX relies on high level protocols such as SPX (Sequential Packet Exchange) to provide a reliable data flow service in sequence. SPX is the protocol of the Novell Transport level and comes from the protocol SSP of Xerox. SPX provides a virtual service, reliable, oriented to the connection between network stations using the IPX service to provide a data flow in sequence. The result is reached by implementing a conversation protocol, requiring the reception of each sent frame to be confirmed before sending the next frame. SPX, besides, provides a flow control between stations and ensures no doubled frame is delivered.

- 103 -

11. UPPER LEVEL PROTOCOLS

Each system connected to a network IPX/SPX is identified by an univocal address in 80-bit format. The first 32 bits of the address coincide with the network identification range (network), while the 48 next bits identify the host. The network identification range detects the single network on which the source or destination system are located, while the host identification range identifies a specific system of that network, and usually coincides with the address of the NIC card of the system.

Fig.11.3 NetWare Architecture

Fig.11.4 NetWare Architecture

- 104 -

11. UPPER LEVEL PROTOCOLS

11.3 The NETBIOS/NETBEUI PROTOCOL The term NetBIOS (Network Basic Input/Output System) concerns the services network extension that have been historically offered by the BIOS (Basic I/O System), software resident in the ROM of each IBMcompatibile PC. Actually the NetBIOS extends the I/O capacities of a PC to the network ambient, defining an API interface (Application Program Interface) for the request of communication services of the Transport level and a protocol describing the way in which these services are provided. NetBIOS has become a standard "de facto" of the Session level in the LAN communication, and the NetBIOS services are offered by a large range of network software. The name NetBEUI (NETBios Extended User Interface) is sometimes used to refer to the protocol supporting the NetBIOS services and/or the software components used to implement it. NetBEUI was the name of the software components developed at first by IBM and Microsoft to support the NetBIOS services of the first PCs. Unlike the other protocols (TCP/IP and IPX/SPX), NetBIOS identifies the users of a network according to an alphanumeric name, and not according to a specific address. In NetBIOS the names represent the base of communication between application programs. NetBIOS has a table for each system identifying the name or the names associated to the application programs of that system, enabling a given application program to have more names, and a given system to support more user names simultaneously. The names can be added or cancelled dynamically and the same name can be given to different systems in different times: two systems with the same name cannot be active on the LAN in the same instant. The name can be a single name or a group name: the single name identifies the single system, while the group names enable different users on the same system or on different systems to be addressed using only one name (transmission in multicast). NetBIOS associates the names directly to the station addresses and does not determine any separate addressing mechanism in the Network level, unlike TCP/IP and IPX/SPX. For this reason, the use of the Transport protocol NetBIOS does not fit an internet composed by more LAN interconnected by means of router, as the router does not have - at Network level - the information necessary to decide toward which networks to route the received frames. NetBIOS includes two data distribution services related to the Transport levels, a datagram service without connection and a service with - 105 -

11. UPPER LEVEL PROTOCOLS

connection often called session service, beside a set of other support services for the control of the LAN communications.

Fig.11.5 Frame NetBIOS

- 106 -

11. UPPER LEVEL PROTOCOLS

11.4 QUESTIONS

Q1

Who did define the IPX/SPX protocol ? 1 2 3 4

Q2

Which service does the IPX protocol provide? 1 2 3 4

Q3

2 3 4

An additional EPROM changing the PC BIOS to support the network applications A software interface to the high level or session API application programs A protocol to connect the physical layer to the logical layer in the OSI model A control method for the access to the transmission mean

Which of the following protocols is used the most in the Internet network ? 1 2 3 4

Q5

To deliver a packet to the destination, without acknowledging receipt To control the flow between two or more stations To convert an IPX packet into TCP/IP To deliver a packet to the destination, acknowledging receipt

What is the NetBIOS ? 1

Q4

Xerox IBM Microsoft Novell

IPX/SPX TCP/IP NetBIOS Others

Which of the following protocols is the less flexible to join networks of different kinds ? 1 2 3 4

IPX/SPX TCP/IP NetBIOS NetBEUI

- 107 -

12. CONFIGURATION of a LOCAL AREA NETWORK

12. CONFIGURATION of a LOCAL AREA NETWORK 12.1 DESIGNING a LAN A PC can be easily changed, but to change the asset of a local area network needs strong investments of money and time, and the acquisition of new technical notions, too. To carry out a local area network, so, is a long term investment, a target to be centered from the beginning even to the detriment of an expensive analysis and design. As any other computer science system, its structure combines heterogeneous elements which must be chosen especially for efficiency and flexibility of use, not considering the costs much, at least at the beginning. A network device which does not operate or operates badly, compromises the whole LAN completely or partly frustrating the time and the investment necessary to its preparation. 12.1.1 Choice of the cabling system In small local area networks, it is possible to use a coaxial cable in order to carry out a LAN with bus topology. In this case, the network constitution is very economic as it prevents the use of devices such as the Hub or Switch, but a fault in any point is sufficient to make unusable the whole segment employed by the users. Nowadays, the not shielded twisted pair (UTP) is the most indicated solution for any kind of plant remaining within a single building. The most used twisted pair is the category 5, that, within a certain distance, can carry data even at a speed of 155 Mb/s. The advantage of the twisted pair is to be economic, easy to install, and matching any kind of traffic: local area network, terminal emulation telephony. Due to the low cost of connection, it is usually cabled in excess, covering also those work areas not active at the moment. In this way, whenever you move, it will be enough to activate the nearest access turret changing the interconnections in the central floor or area closet (concentrator). When you need to interconnect the different closets present in separate buildings or on different floors of the same building, the optical fiber is preferred, as it can stand longer lengths, as well as for its high immunity to electromagnetic disturbances.

- 108 -

12. CONFIGURATION of a LOCAL AREA NETWORK

12.1.2 Maximum network extension The maximum extension of a local area network depends on many factors, such as the transit speed of the information, the used standard, the use of Hub, Switch, Bridge, etc, and the kind of used cabling. 12.1.3 Ethernet 10Base5 In an Ethernet 10Base5 local area network, a coaxial cable is used with thick section (about 8 mm. of diameter, RG8), with which it is possible to carry out a segment of 500m maximum total length. In the segment, up to 100 MDI devices (Medium Dependent Interface) or vampire/N transceivers can be connected. The segment is always closed by two 50Ohm terminations. Each MDI is connected to the terminals via the dropcable AUI/AUI. The minimum distance between two MDIs is 2.5 meters. It is possible to connect up to 5 segments in cascade using 4 Repeaters, for a total of 2500m of maximum diameter. Maximum length of the single segment Maximum network diameter Minimum distance between two MDIs Maximum number of Repeaters between two systems Maximum number of MDI per segment

Fig.12.1 Example of Ethernet 10Base5 network

- 109 -

500 m 2500 m 2.5 m 4 100

12. CONFIGURATION of a LOCAL AREA NETWORK

12.1.4 Ethernet 10Base2 In an Ethernet 10Base2 local area network, a coaxial cable is used with thin section (about 5 mm. of diameter, RG58), with which it is possible to carry out a segment of 185m maximum total length. In the segment, up to 30 systems can be connected by means of a passing connection using a special "T"-shaped joint with bayonet connection (BNC), to be connected to the BNC of the NIC card. The segment is always closed by two 50-Ohm terminations. The minimum distance between two systems is 0.5 meters. It is possible to connect up to 5 segments in cascade using 4 Repeaters, for a total of 925m of maximum diameter. Maximum length of the single segment Maximum diameter of the network Minimum distance between two MDIs Maximum number of repeaters between two systems Maximum number of systems per segment

Fig.12.2 Example of Ethernet 10Base2 network

- 110 -

185 m 925 m 0.5 m 4 30

12. CONFIGURATION of a LOCAL AREA NETWORK

12.1.5 Ethernet 10BaseT In an Ethernet 10BaseT local area network, a multipair cable is used, with which it is possible to carry out segments of maximum length equal to 100m. This limit is set by the standards for the structured cabling system. It is possible to connect up to 5 segments in cascade using 4 Repeaters (most commonly known as Hub), for a total of 500 meters of maximum diameter. Maximum length of the single segment Maximum network diameter Maximum number of Hubs between two systems Maximum number of systems per segment

Fig.12.3 Example of Ethernet 10BaseT network

- 111 -

100 m 500 m 4 2

12. CONFIGURATION of a LOCAL AREA NETWORK

12.1.6 Ethernet 10BaseFB/FL In an Ethernet 10BaseFB/FL local area network, an optical fiber cable is used, with which it is possible to carry out segments of 2000m maximum length. It is possible to connect up to 5 segments in cascade using 4 Repeaters or optical Hubs, for a maximum diameter variable as function of the IPG (Inter Packet Gap) and the RTD (Round Trip Delay), and however usually limited to the maximum segment extension. Maximum length of the single segment Maximum network diameter Max. number of repeaters/Hubs between 2 systems Maximum number of systems per segment

Fig.12.4 Example of Ethernet 10BaseFB/FL network

- 112 -

2000 m variable 4 2

12. CONFIGURATION of a LOCAL AREA NETWORK

12.1.7 Ethernet 100BaseTX In a Fast-Ethernet 100BaseTX local area network, a multipair cable is used, with which it is possible to carry out segments of 100m maximum length. This limit is set by the standards for the structured cabling system. It is possible to connect up to 2 segments in cascade using 2 Repeaters or Hubs, for a maximum diameter of 205 meters. Whenever a higher distance between the Hubs should be necessary, the total distance between two systems must never be over 205m. Maximum length of the single segment Maximum network diameter Max. number of repeaters/Hubs between 2 systems Maximum number of systems per segment

100 m 205 m 2 2

12.1.8 Ethernet 100BaseFX In a local area network Fast-Ethernet 100BaseFX, an optical fiber cable is used, with which it is possible to carry out segments of 412m maximum length in "half-duplex" mode and of 2000m in "full-duplex" mode. It is possible to connect up to 5 segments in cascade using 4 Repeaters or Hubs, for a maximum diameter variable as function of the IPG (Inter Packet Gap) and the RTD (Round Trip Delay), and however usually limited to the maximum segment extension. Maximum length of the single segment (HD) Maximum length of the single segment (FD) Maximum network diameter Max. number of repeaters/Hubs between 2 systems Maximum number of systems per segment

Fig.12.5 Example of Fast-Ethernet 100BaseTX network

- 113 -

412 m 2000 m variable 4 2

12. CONFIGURATION of a LOCAL AREA NETWORK

12.2 MIXED CONFIGURATIONS We have treated the transport means and the equipment which can carry out and interconnect different cabling and LAN architectures at physical level. Now we want to deal with and describe the aspects enabling the interconnection of different types of networks using different physical levels and protocols. • •

•

A first aspect concerns the evaluation which must be made in terms of network design or change concerning its extension A second aspect concerns the kind of protocol used in the different network segments and so the used NOS ambient (Network Operating System) A third aspect concerns the division of the users belonging to the network into work groups, and the kind of software they use for their activity.

12.2.1 The transmission mean and the network technology Many of the operations of the installed networks depend on the physical level as described before. The designer and the installer must have a detailed plant of the building to expect and quantify the distances between single workstations and server. In calculating the distances, not only the distance in a bee-line between the single systems must be counted, but also the vertical trunks from floor to ceiling toward the different workstations: i.e. the effective development of the transmission mean is calculated. The paths chosen to set the transmission mean must not be adjacent to the electrical conductors of the building, but at a minimum distance fixed by the rules. The systems located in the same area are almost always cabled with STP/UTP cable up to a typical distance of 100m between user’s station and competent administration closet. This gives the possibility to insulate a faulty system with ease, with the advantage to easily switch from the speed of 10Mb/s to the higher one of 100Mb/s. The center of this area normally consists of the Hub which cabling ray will be 100m. Very often network equipment are installed such as Hub, Switch and Router inside closets which can group their operations, although this solution cannot always be applied if the distance of some stations overcomes 100m. A solution can be to decentralize the Hubs inside the building or to use Repeaters in longer trunks. A second solution expects to change the kind of cable to join the two quantities set at a superior distance. A 10Base2 coaxial cable can join a pair of Hubs or single stations at a distance of about 200m, the 10Base5 cable up to about 500m. This solution implies only a variation of the kind of cabling, but includes also the addition of further accessories such as transceivers or signal converters. For this reasons, many Hubs include one or more ports with - 114 -

12. CONFIGURATION of a LOCAL AREA NETWORK

different standards, to solve the problem in the most economic and fast way (BNC and AUI ports). Consider that the use of coaxial cable trunks is good only for speeds of 10 Mb/s. A valid but expensive solution can be the optical fiber, which can support the standard for 10/100 Mb/s. The fiber is used especially in extremely long trunks set on critical paths. Among the most interesting cases, there are the underground lines which can join two or more buildings belonging to a single interbuilding. The underground lines are exposed to the danger of extremely variable electrical potentials, according to the conditions of the soil and the meteorological conditions (lightning). It is advisable to use this kind of cabling in these situations. Particular equipment such as the Router and the Switch tends to be used to split work groups between them so not to jam the network with repeated messages to users not using this information. For this reason, these systems can be grouped into closets so to facilitate the modification of the network structure acting from a single point. Remember, that these equipment can have a certain number of LAN/WAN ports configurable at wish by the network supervisor, in order to create particular pseudo-dedicated connections and optimize the traffic of the LAN packets. Using the solutions suggested by a single manufacturer there is the great advantage to use proprietary buses to interconnect Router, Hub and Switch. In fact, on the rear of these devices there are often flat-cable connectors which must be used to connect all homogenous devices in parallel. These connections do not affect the existing cabling at all and do not belong to it, but constitute the special buses to exchange the data among these devices at very high speed, without using the expensive optical fibers. We will now speak of stackable equipment, which can be connected in a stack. Switching to the NIC network cards, they can be bought in the "combo" version so to match to the different cabling solutions. Usually a "combo" is a card provided with BNC (10Base2), AUI (10Base5 for connection with transceiver) and RJ45 connection (10BaseT).

- 115 -

12. CONFIGURATION of a LOCAL AREA NETWORK

12.2.2 The Communication Protocol The protocol has really an important role in the connection of different networks, and it must be chosen as function of the kind of machines and operative systems to be supported. The suggested rule includes the use of a single kind of protocol on the carried out network. This rule, however, cannot always be respected, as many operative systems have proprietary protocols and operate at best only with these. Remember that in an existing cabled network, frames can flow using different protocols without causing anomalies. Only systems using the same protocol can communicate between them inside a network: in this way, the frames of different protocols will be simply ignored. E.g., consider a building provided with Novell IntranetWare server with IPX/SPX protocol and a Microsoft Windows NT Server 4.0 server with TCP/IP and/or NetBEUI protocol. The work-stations provided with IPX/SPX protocol can access the resources available only on the Novell server, as well as those provided with TCP/IP and/or NetBEUI protocol can access only the Microsoft server. Using the same protocols on the stations you can have full access to all network resources. This last solution, although possible is not valid anymore, as it increases the work of the single stations decreasing their performances. For this reason, many network software developers provide different protocols supports on the same network operative system, so that it is the same server that carries out the protocol conversion. In the Intranet networks, it is almost necessary to use the TCP/IP protocol to favor the use of applications using this standard. 12.2.3 Homogeneous grouping of the users The division into work groups is another main point in designing a network. By properly dividing the users as function of the purposes and their actual necessities on the resource sharing, you can optimize the LAN traffic and often disengage too congested segments. Whenever particular users should result as medium points of work groups, it is possible to give them their privileged connections directly from the Switch or the Router. The users can be divided by functional area (Administrator, Secretariat, Storehouse, Production, etc.) or as function of the employed user’s operative system (Windows, OS2, etc.) and the kind of server they access (Novell IntranetWare, Microsoft NT Server, IBM AS/400, etc.).

- 116 -

12. CONFIGURATION of a LOCAL AREA NETWORK

12.3 QUESTIONS

Q1

What must be considered in designing an efficient network ? 1 2 3 4

Q2

In order to obtain a good cabling system, what must be considered ? 1 2 3 4

Q3

25 meters 50 meters 100 meters 186 meters

In a Fast-Ethernet 100BaseTX network, what is the maximum length of the trunk between station and concentrator ? 1 2 3 4

Q5

The number of terminals and the network diameter The technology using the really fastest support The cost of the support and the connectors The design enabling the easiest maintenance

In an Ethernet 10BaseT network, which is the maximum length of the trunk between station and concentrator? 1 2 3 4

Q4

The minimum number of terminals in order to reshape and redesign the whole set in a second time The number of hubs and switches to be used The number of connection properly oversized, the speed and the ambient where it will be installed The cost

25 meters 500 meters 100 meters 186 meters

Using the standard Fast-Ethernet 100BaseTX, how many concentrators are allowed ? 1 2 3 4

one two four none

- 117 -

13. NETWORK CONTROL and ADMINISTRATION

13. NETWORK CONTROL AND ADMINISTRATION This chapter describes the techniques and instruments for networks control. Three levels will be treated: • • •

Control and introduction of the network information Protocols analysis and traffic control Statistic analysis

13.1.1 Control and introduction of network information The software packets for network control must introduce the alarm situations and record intermittent events and malfunctions present inside the system. The ISO is suggesting a new architecture indicated as CMIP (Common Management Information Protocol) performing the above functions. The functions fixed by the model include the control of the errors, the configurations, the performances, the safety and the accounting. The error control includes the fault detection and their insulation. The configuration control provides messages describing the machines and active connections, and is strictly connected to the error control, as the modification of the configurations is the main technique to insulate errors from the network. The performance control counts the frames, the access requests to the disks and to the specific programs. The safety control includes the warning to those performing not authorized access attempts to the transmission mean, to the network, the server and generally the common resources. The accounting control concerns the generation of the reports sent to the users on their use of the resources. An alternative to CMIP is represented by the SNMP (Simple Network Management Protocol). The Simple Management Protocol has been designed to enable the communication between the software present on a Network Management Station (NMS) and the software agents on the controlled devices. The communications consist in cyclic interrogations generated by the NMS toward the network equipment, to which they reply with error signaling messages and/or statistics by means of the above mentioned software agents. - 118 -

13. NETWORK CONTROL and ADMINISTRATION

The cons of the SNMP are the lack of safety, the scarce quality of the documentation and the tendency of some firms to create non standard configurations. The SNMP is to be used with networks with TCP/IP protocol, but usually not with networks with IPX/SPX or NetBEUI protocol. In spite of these problems, the industry continues to use this system which does not occupy too much memory and capacity of calculation of the equipment and the NMS. There are other products such as the IBM NetView having the pro to be powerful and reliable, although expensive, and the con to need high calculation capacities on the system housing it. Other competing systems such as the Novell, Microsoft and Banyan offer, however, a set of utilities dedicated to their servers and to the filesystem they control. 13.1.2 Protocols analysis and traffic control A further help for a network control is the probe, a system provided with local microprocessor which the connections to the different network sub-systems or to the different network segments branch from. In practice, through these connections, the probe can monitor the physical and logical parameters of the network at lower costs than a protocol analyzer. The probe is often connected to the PC via serial port or to the network via LAN interface. In this case, by means of the specific software RMON (Remote Monitoring) on board, it will collect the statistic data on the monitored LAN activity, keeping them in memory and making them available when required by an NMS, actually behaving like a network/protocol analyzer and an NMS. An expensive but complete alternative is the protocol analyzer. These are instruments for network analysis that can support any protocol and any kind of interface. These products are similar to portable PC including the hardware and software for their interconnection to the existing cabling. Most analyzers contain a TDR (Time Domain Reflectometer) that can detect cut-off cables, measure BER (bit error rate), analyze the information flow between two systems, carry out protocol statistics, etc.

- 119 -

13. NETWORK CONTROL and ADMINISTRATION

13.1.3 Statistical analysis The networks are dynamic systems: their operation can be defined in terms of measurable parameters, used by those in charge of growth planning, for fixing the comparison terms, for detecting the problems in the starting phases and justifying the estimates. On this purpose, many are the products available to provide those in charge of the LAN with statistical data, in common or quantified shape as concerns the network equipment as well as its centralized systems. By carefully analyzing these data, efficient and productive network environments can be created. The products generating statistical relations are usually software completing the capacities the system is already provided with, and usually providing information such as: • • • • • • •

Loads of the LAN divided by protocol, users and time Errors of the LAN divided by type and concentration period Space and memory on disk used by particular applications, users or cost centers Quantity of activities in specific programs or files Duration of the connections of specific users or workstations Number of print works (expressed in different ways) Quantity of work on the server in given periods of time.

Some programs prevent potential problems on the network, as for example the infection from virus, controlling the symptoms and warning the one in charge of the LAN in case non authorized users are detected, or if they try to enter the network. Others inform the one in charge if there is network jamming, improving the efficiency and performances in this way. Different programs signal if the server needs supplementary memory or if a software application must be up-dated. In the end, the statistical relation packets are not particularly attractive, but useful for the person in charge to ensure a higher efficiency, a better planning of the activities and detection of the potentially dangerous situations before they become problems.

- 120 -

13. NETWORK CONTROL and ADMINISTRATION

13.4 QUESTIONS

Q1

The Simple Management Protocol has been designed to : 1 2 3 4

Q2

The probe is an instrument to : 1 2 3 4

Q3

4

A person choosing the best protocol for the existing network An instrument for monitoring all network analytically An instrument for reading and coding of the packets crossing the network An instrument counting the packets sent back because wrong

What is the use of a TDR in a network ambient ? 1 2 3 4

Q5

Check the operation of the single trunk Check the connected stations only See the sources of the computer connected to the network Monitor the traffic, the physical and routing parameters of the network

What is the protocol analyzer? 1 2 3

Q4

Easily control the change of protocol inside the network Help the system’s Administrators to monitor the traffic in the network Change the access priorities to the different stations to optimize the network performances Enable the communication between the software on the network control station and the agents on the controlled devices

To send a packet to all stations when the system’s administrator needs it To measure the network diameter to check if the max. limit has been reached To detect malfunctioning trunks To measure the real speed of the traffic in the network by controlling the number of collisions

Which purpose does the network Administrator use the statistic for ? 1 2 3 4

To optimize the network, plan the growth and detect the problems To measure the productivity of the single terminals To detect the less used trunks to exploit the band To justify the maintenance costs

- 121 -

14. NETWORK OPERATING SYSTEMS - NOS

14. NETWORK OPERATING SYSTEMS 14.1 WHAT is a NOS The purpose of a network operating system is to control and conform the operations the users can perform on a local area network. Basically there are many needs to be controlled inside a work group: • • • • • • • •

the possibility to memorize the data in a single file so to make them available to anyone the inner control of safety for such database, so to make them available only and exclusively to those having the right the possibility to share the resources (see peripherals) present only on some systems belonging to the work group the possibility to connect the existing network to a new network or to expand the one already existing the control of safety of the data memorized in the server via backup copies the access to the network using different operative systems or different locations (included the remote access via modem) the control of the generation and forwarding of the electronic mail on the existing network or on Internet the construction of hierarchical structures inside the work group so to subdivide responsibilities and manage the work into classes.

A network operative system is usually installed on a dedicated machine called Server, which purpose is to satisfy the above described requirements. On the market, at the moment there are many packets for this purpose. In particular the characteristics of the Novell IntranetWare and MSWindows NT Server will be described.

- 122 -

14. NETWORK OPERATING SYSTEMS - NOS

14.2 NOVELL INTRANETWARE This operative system includes the activation of a dedicated Server provided with NFS partitions (Netware File System) containing the services and the database, the users’ authentication, the print control system and other network services. NFS has been designed for a file sharing with high performances in data control and safety. NFS organizes the memorization of the information on disk into volumes, directories, sub-directories and files. The users are granted rights of access and use for specific directories of the server. Besides the rights of access, IntranetWare includes a set of devices to protect the data from physical damages. They can be used during the blocking phase of a record or a range and, whenever the modification process should not be successful, it is always possible to reset the last situation before blocking. Another important IntranetWare characteristic is the NDS (NetWare Directories Services), with which the supervisor can control very big networks (Enterprise Network) in a reliable and quick way. The system user interface of the operative system is proprietary and is organized via a menu. Most of the control activities of the Server must be carried out by a user system, as the Server enables a restricted number of operations: the supervisor can administrate the network from any station using proper utilities. Each user employs a proper client program operating under MSWindows/DOS, Apple Mac or IBM OS/2 to access the only resources of the Server. In practice, no user can interact or directly access the resources of other machines (excluded the single printers). This provides a high safety always regulated by the username/password mechanisms. Each user is provided with rights or privileges granted by the network supervisor. The prints can be controlled via printer-server services activated on the same Server, on a user system or on an independent system connected to the LAN. The network protocol is the proprietary IPX/SPX one, although in the last version it is possible to pair it to the TCP/IP protocol to build up Intranet networks, from which the name IntranetWare. The Server becomes the center of the same network, so it must be protected with the usual devices (E.L.C.B., backup system, etc.). The Server supports also advanced Routing and Bridging functions included in the operative system.

- 123 -

14. NETWORK OPERATING SYSTEMS - NOS

14.3 MICROSOFT WINDOWS NT SERVER This operative system includes the activation of a non-dedicated network Server. It is similar to the last as concerns the characteristics, although showing some improvement in terms of sources control which make it the best tool for small/average users. The software must be installed in a machine called Server with the same characteristics of the one seen for the Novell ambient. This operative system includes the activation of a non dedicated Server provided with 16-bits FAT or 32-bit NTFS (NT File System) partitions containing the services and database, the user’s authentication, the print control system and the other network services. NTFS has been designed to carry out a file sharing with high performances in the data control and safety, although not comparable to NFS (Netware File System) of Novell still today. As NFS, even NTFS organizes the information memorization on disk in volume, directory, sub-directory and file. The users are given access and use rights to specific directories of the Server. Besides the access rights, Windows NT Server includes also a set of expedients to protect the data from damages of physical kind in a very similar way as included in IntranetWare. Unlike IntranetWare, Windows NT Server uses a domain control for very big networks, through which the supervisor can control them very quickly. The user interface of the operative system is very similar to the interface of Windows 95, which eases the activity of the supervisor. All control activities of the Server can be carried out on the same Server, although the main control activities can be assigned to a remote user’s system. To access the Server each user can use the same network services included for Windows 95. In practice, each user can interact or access directly the server and other machines resources. This includes a high degree of flexibility, although keeping a safety level always adjusted by the username/password mechanism. Each user is provided with rights or privileges granted by the network supervisor. The prints can be controlled with printer-server services activated on the same Server, on a user’s system or on a independent system connected in LAN. The network protocol is the proprietary NetBEUI one, but it is possible to help it with the protocol IPX/SPX or TCP/IP to carry out Intranet networks. The Server becomes, so, the center of the same network, so it must be protected with the usual devices (E.L.C.B., backup system, etc.). The server supports also some routing functions included in the operative system.

- 124 -

14. NETWORK OPERATING SYSTEMS - NOS

14.4 QUESTIONS

Q1

What is a NOS ? 1 2 3 4

Q2

If you must privilege the resource sharing on the workstation, which operative system would you chose ? 1 2 3 4

Q3

The owner of the firm The backup user The network supervisor All users

Which of the following systems has a good level graphical interface ? 1 2 3 4

Q6

NTFS FAT NFS HPFS

Who in a factory will be in complete charge of the NOS ? 1 2 3 4

Q5

Novell IntranetWare. Microsoft Windows NT Server. Microsoft Windows 95. Microsoft MS-DOS.

Which of the following file systems is used by Novell ? 1 2 3 4

Q4

A procedure operating on workstation The operating system used on the server The driver of the network card A backup program for the data

Novell IntranetWare Microsoft Windows NT Server Microsoft MS-DOS Microsoft Windows 95

What must a network operating system guarantee ? 1 2 3

Database safety, hierarchical organization of the work and resource sharing groups on the server Prints in local and sharing of all stations disks Routing and bridging operations on the packets

- 125 -

15. NETWORK INTERNET / INTRANET

15. INTERNET / INTRANET NETWORK 15.1 INTRODUCTORY CONCEPTS In the first half of the 70ies, the DARPA (Defence Advanced Research Project Agency) started developing a packet switching network for interconnection of heterogeneous calculators, to be used as communication mean between the research institution of the United States. At the end of the 70ies two more famous protocols were defined, the TCP and the IP. These protocols were used by a group of researchers for the ArpaNET network and obtained a great success as available and usable by all. From the 1990 the ISO wanted to standardize the TCP/IP protocol as this was heavily diffused on the world market. Actually, there are many networks permanently interconnected, constituting a single body called Internet. This network, mainly consisting in processors belonging to Universities and Research Centers, can be accessed by anyone if provided with a terminal and a Point of Presence (POP) on the same. Many business organizations offer contracts for the remote access to the network via modems or dedicated lines. These organizations are called Provider. For consultation (navigation), the network provides Servers dedicated to heterogeneous information concerning all possible subjects. Besides, Internet offers a valid communication instrument to send the electronic mail at world level. 15.2 PROTOCOLS and SERVICES Hereafter follows a list of the most important protocols and services used today in Internet. 15.2.1 The IP (Internet Protocol) IP is the Network level protocol for support of an Internet network station. It is a simple protocol that can route the messages on the network and assemble and control the messages generated by the stations. The IP protocol univocally identifies a station inside the network with the use of an address consisting in 4 bytes separated by decimal points (see chapt.11.1). These addresses can be given in static way (via some bodies) or can be experimental, i.e. for small inner networks. In alternative to the address, a station can be detected from its host name: this methodology must be supported with a DNS (Domain Name Server) which can guarantee the name table conversion into IP addresses.

- 126 -

15. NETWORK INTERNET / INTRANET

We can say that the Internet addresses have two formats: one for the computers (expressed in numbers) and one for the persons (expressed in words). 15.2.2 The TCP (Transmission Control Protocol) The TCP is a Transport level protocol used in an Internet network. It provides different characteristics among which the direct support to the application using it. Its main purpose is to provide a reliable communication between stations, always needing acknowledgement of the sent data reception. 15.2.3 Telnet Telnet is a protocol enabling the PC connection to any other remote system connected to the network in terminal emulation, only if the remote system provides and accepts this service. The connection is activated by inserting the remote calculator name or its IP address after the command "telnet". From that moment on, all characters typed on the keyboard will be transferred to the remote and the responses it generates will be displayed on the local screen. The Telnet program usually includes the emulators for the most used terminals (VT100 and ANSI). 15.2.4 FTP (File Transfer Protocol) It is an application enabling a user to transfer files from and to a remote system. The safety levels are controlled by logon procedures (connection) to the service, with username and password use. FTP transfers any file format and can code the received objects into different formats. 15.2.5 TFTP (Trivial File Transfer Protocol) It is a simplified version of FTP, where there is no logon procedure available with username and password use. It is usually employed for control activities of communication systems requiring a file transfer without authentication.

- 127 -

15. NETWORK INTERNET / INTRANET

15.2.6 SMTP (Simple Mail Transfer Protocol) It is one of most important applications of the TCP/IP. It enables to send electronic mail to the network users. Each user is identified by the syntax: UserName@DomainName, and usually no authorization is needed to send an electronic mail message. The sending procedure systematically tries to send more times until it achieves its goal. The message destination user will receive notification of the occurred reception. 15.2.7 DNS (Domain Name Server) The DNS is a data base distributed and replied on different systems. Its purpose is to define the correspondences between host names and IP addresses in real time. This enables a user to select a destination only knowing his name. 15.2.8 SNMP (Simple Network Management Protocol) It is a protocol designed to send data dedicated only to network equipment, such as Router, Bridge, Switch, Hub, etc. In this way a network administrator can control all communication sub-systems from a single work-place. 15.2.9 NIR (Network Information Retrieval) These are spread hyper-textual services enabling the access of a wide range of information in a simple way, ignoring where the information is deposited. Created to enable the use of the Internet network also by less expert people, they have been greatly successful and quickly spread: the most known service of this category is the WWW (World Wide Web). For DOS/Windows compatible systems there are software packets such as the Microsoft Explorer and the Netscape Communicator implementing the service browser.

- 128 -

15. NETWORK INTERNET / INTRANET

15.3 INTRANET NETWORKS After the great success achieved by Internet and the services it provides, recently they thought to use the same technologies for private networks, too, using the TCP/IP protocols and part or all services present in Internet, such as WWW, FTP, Telnet, SMTP, etc. The term Intranet includes the technology enabling to store, process and distribute data of an organization to the users of their own inner networks, using the Internet network technology locally. So, we define as Intranet a local area network installed in a restricted place (building, industry, etc.) using the TCP/IP as standard protocol, and at least a Server with network operative system which can make the above mentioned protocols and services available. The advantage of this solution concerns the singularity of the used protocol and the possibility to connect the inner network to Internet using only a Router. The last has the function to control the inner safety so to prevent not wished accesses from the outside. Besides, with these networks you can easily control the electronic mail services and those for document availability. Inside an Intranet network, each station can be identified by a host name and an IP address. Almost always, the choice of the address locates on the experimental addresses set (192.168.x.x), so not to cause malfunctioning on the packets routing processes whenever a direct connection to Internet should be activated in the future.

- 129 -

15. NETWORK INTERNET / INTRANET

15.4 QUESTIONS

Q1

In which way does the IP protocol univocally identify a station inside the network ? 1 2 3 4

Q2

What does the Telnet protocol enable ? 1 2 3 4

Q3

Transfer file in Internet Send and receive electronic mail Have your own Web page Chat in Internet with other users

What does NIR mean ? 1 2 3 4

Q6

Set the files of your own terminal available Unload files in Internet Transfer files from and toward a remote processor Chat with other users connected in Internet

What does the SMTP protocol enable to? 1 2 3 4

Q5

Carry out file transfer on geographical networks Chat with messages in the network Connect to the server with the Telnet user_name service Connect to any remote system connected to the network

What is the use of FTP ? 1 2 3 4

Q4

Four bytes separated by decimal points The user login name The station physical position The login name and the user password.

A protocol for data transmission in Internet A system for identification of the users connected in Internet A browser to display the Web pages A mail messages display

What is an Intranet network? 1 2 3 4

A network on which all users can use the electronic mail A set of limited servers, all connected to Internet A business network which is connected to Internet via a switch A private network using Internet protocols and services

- 130 -