Topic2 Network

Architecture

LEARNING OUTCOMES
Bytheendofthistopic,youshouldbeableto: 1. 2. 3. 4. Describe three basic types of topologies, Explain the benefits and uses of different topologies. Discuss the Ethernet and Token Ring network transport systems. Describe the features of different network devices

INTRODUCTION
This topic describes the basic layout of networks. It also covers the two most commonly used transport systems: Ethernet and Token Ring. Once you learn the fundamentals of network architecture, you should be able to design a network in a high level.

2.1

SIMPLE LAN TOPOLOGIES

Network topology is the way nodes are arranged and connected with links. Along with the physical medium chosen for implementing the links, it determines the speed of the network and the communication efficiency. Its selection depends on the geographic environment, the kinds of application running on the network and the implementation costs. As discussed previously, three types of topology are commonly used: bus, star and ring.

26

TOPIC 2

NETWORK ARCHITECTURE

(a)

Bus Nodes may be distributed in an area like an office or a campus. In the bus topology, a single link (the bus) runs across the area, and nodes can be tapped to it at any point. This topology should be implemented with a transmission medium that allows multipoint access. To ensure that no signals bounce back at the ends of the bus and interfere with the trailing signal, terminating devices are plugged at the ends. Refer to Figure 2(a) for the configuration of the bus topology. Suppose node A has a message to send to node B. Node A delivers the message to the network via its network adaptor. The message propagates in both directions of the bus until the ends are reached. All nodes tapped to the bus can receive the message but only node B, which finds itself as the destination of the message, copies the message into its memory. You should note that node B only copies the message from the network, not remove it. At the end points of the bus, the terminating devices absorb the signals to clear up the link.
Nodes Frame

Terminator

Link

Figure 2(a): Bus topology

Headend

Figure 2(b): Tree topology

TOPIC 2

NETWORK ARCHITECTURE

27

Though the bus is shared by multiple nodes, only one signal can be transmitted at one time over the network. If two nodes send out messages at the same time, a clash of the signals in the link results. Some form of medium access control is needed to manage the use of the common link. We will have a more detailed look into this issue in the section on network technologies, further on in this topic. A topology that extends from the bus topology is the tree topology. It is a bus topology with branches, as shown in Figure 2(b). Its underlying mechanism is the same as that in the bus topology, except that it has more than two end points. We call the end where all the branches concentrate, the headend. (b) Ring In contrast to the single link (the bus) with open ends in the bus topology, nodes in the ring topology are arranged to form a single closed loop. Actually, each node is connected to the network via a repeater and the repeaters are connected by point-to-point links to form a ring configuration. Refer to Figure 2.1 for the ring topology.
Repeater

flow of data

Figure 2.1: Ring topology

Each repeater has three connections: one to the node, and the other two to two separate links connecting to another repeater at the other end. The repeater is responsible for receiving data from one link and immediately transmitting it, bit by bit, to the other link. If the repeater finds the message is destined to the node it is attached to, the message is copied into the memory of the node. As there are no end points in the ring topology, you may wonder how the signals on the link are cleared. A message sent out from a node continues to circulate from one node to another until the source node is reached again. The source then absorbs the message, thus clearing it from the network. The links used are unidirectional. That is, the

28

TOPIC 2

NETWORK ARCHITECTURE

signals transmitted in the ring flow only in one direction, either clockwise or anti-clockwise. (c) Star The star topology is a central node to which all the other nodes are connected, resulting in a star configuration as shown in Figure 2.2. Each node connects to the central node with a point-to-point link. A message sent out from a node is first received by the central node. If it is not destined to the central node, it is redirected to all outgoing links attached to the central node. All other nodes can then receive the message and the destination node copies the message into its memory.

Hub

Figure 2.2: Star topology

The central node is often a device called a hub. A hub is used to concentrate all of the links into a single point in the network and it usually has multiple ports for multiple links to plug into. Hubs are further described in the section on network hardware components. (d) Hybrid LAN Topology Except in very small networks, you will rarely encounter a network that follows a strict bus, ring or star topology. Simple topologies are too restrictive, particularly if the LAN must accommodate a large number of devices. More likely, you will work with a complex combination of these topologies, known as a hybrid topology. (e) Star Bus Topology Figure 2.3 shows the star bus topology. Instead of a single central hub, nodes in a network are connected to several hubs and there are links connected between the hubs to form a bus trunk. If one node goes down,

TOPIC 2

NETWORK ARCHITECTURE

29

the other nodes will not be affected. If one hub goes down, none of the nodes attached to it can communicate with the network and the bus trunk is also broken.
HUB HUB HUB

Figure 2.3: Star bus topology

SELF-TEST 2.1
What are the primary advantages of using a star topology over a ring or bus topology?

2.2

NETWORK TRANSPORT SYSTEMS

A network transport system is sometimes referred to as the networks logical topology, because it describes the networks logical interconnection between nodes rather than its physical interconnection. The network transport system depends on electrical pulses carried by the physical layer of the OSI model. The two most popular network transport systems are Ethernet and Token Ring. In this section, we introduce the concepts of switching, shared Ethernet and switched Ethernet. Then, we explain both Ethernet and Token Ring systems in more detail.

2.2.1

Switching

Switching is a component of a networks logical topology that determines how connections are created between nodes. There are three methods for switching. There are: Circuit switching Message switching Packet switching

Every network transport system relies on one of these switching mechanisms.

30

TOPIC 2

NETWORK ARCHITECTURE

The next reading introduces the three switching methods. You should note that every network transport system relies on one of these switching mechanisms.

Reading
Dean, pp. 19799.

2.2.2

Shared Ethernet

Ethernet is a networking transport method originally developed by Xerox in the 1970s and later improved by Xerox, Digital Equipment Corporation (DEC) and Intel. This flexible technology can run on a variety of network media and offers excellent throughput at a reasonable cost. Ethernet is, by far, the most popular logical topology for LANs today and its popularity continues to grow. Ethernet follows a set of communications rules called Carrier Sense Multiple Access with Collision Detection (CSMA/CD). All Ethernet networks, independent of their speed or frame type, use CSMA/CD. In Ethernet networks, the network adaptor of each node always monitors the link to see if another node is transmitting data on it. This is actually the meaning of Carrier Sense (CS) in CSMA/CD. All nodes attached to the link know about the state of the link; that is, whether it is idle or being used by someone. A node transmits data to the link immediately when it finds the link is free, as shown in Figure 2.4 Multiple Access (MA) means that it is a technology used in multiple access network. There may be more than one node that wants to transmit data finding the link idle and submit data into the link at the same time. When this happens, a collision occurs. Collision Detect (CD) means that nodes are able to detect collisions during data transmission. If a collision is detected, the node transmits a jamming signal and immediately stops transmitting. This ensures that all nodes on the link know there has been a collision. Each sending node then waits for a random unit of time before retransmitting, in order to reduce the probability of colliding again. If there is further collision on retransmission, the sender doubles the waiting time for retransmission until reaching a limit.

TOPIC 2

NETWORK ARCHITECTURE

31

Figure 2.4: Computer transmits data when the cable is free

Ethernet comes in a variety of implementations. Each Ethernet version follows a slightly different IEEE 802.3 specification that outlines its speed, topology and cabling characteristics. The following section describes the most common Ethernet Physical layer varieties: 10Base2: This is an Ethernet adaptation that, according to IEEE 802.3 standards, uses thin coaxial cable and a simple bus topology. 10Base2 is also called Thinnet or Thin Ethernet. 10Base5: This is the original cabling standard for Ethernet and uses thick, rather than thin, coaxial cable. It is also known as Thicknet or Thick Ethernet. 10BaseT: This is the most ubiquitous Ethernet cabling specification in use today. It uses twisted-pair cabling (the source of the letter T in its name) and a star topology to transmit data. 100BaseX: The high-bandwidth demands of many modern applications, such as live videoconferencing and streaming audio, have created a need for speed, and many networks require more throughputs than is possible with 10-Mbps Ethernet. This is where 100 BaseX, also called Fast Ethernet, comes into play. 100BaseX comes in several different flavours. It can be implemented over 4-pair Cat. 3, 4 or 5 UTP (100BaseT), over 2-pair Cat. 5 UTP or STP (100BaseTX) or over 2-strand fibre-optic cable (100BaseFX). 1000BaseT: This is also called Gigabit Ethernet. This architecture supports data transfer rates of 1Gbps.

32

TOPIC 2

NETWORK ARCHITECTURE

The following table shows a summary of Ethernet specifications:

Table 2: Ethernet Specifications Summary 10Base2 Media 10Base5 10BaseT 100BaseTX 100BaseFX 62.2/125 micron multimode fibre Duplex media interface connector (MIC) ST 412m Point-to-point 100Mbps

Thin coax Thick coax RG-58A/U RG-8 or RG11 BNC connector AUI/DIX (to transceiver)

TIA/EIA TIA/EIA UTP Cat. 5 UTP Cat. 5 RJ-45 modular RJ-45 modular

Connector type

Max. segment length Topology Transfer rate

185m Bus 10Mbps

500m Bus 10Mbps

100m Star 10Mbps

100m Star 100Mbps

2.2.3

Switched Ethernet

Traditional Ethernet LANs (shared Ethernet) supply a fixed amount of bandwidth that must be shared by all devices on a segment. Stations cannot send and receive data simultaneously, nor can they transmit a signal when another station on their segment is sending or receiving data. In contrast, a switch is a device that can separate a network segment into smaller segments, with each segment independent of the others and supporting its own traffic. Switched Ethernet is a newer Ethernet model that enables multiple nodes to simultaneously transmit and receive data and to individually take advantage of more bandwidth because they are assigned separate logical network segments through switching.

2.2.4

Ethernet (CSMA/CD)

As described in the previous section, Ethernet networks use the CSMA/CD technique for data transmission. Obviously, a considerable amount of bandwidth is wasted in collisions. Consider the worst situation: A collision occurs between two nodes A and B at the two ends of the link as shown in Figure 2.5. We assume the end-to-end propagation delay to be a. At time t0, node A finds the link idle and transmits data. Node B still finds the link idle until it receives the first bit of the frame from node A. If B starts transmitting data at the time just before the frame from node A arrives (at a time near t0 + a), a collision occurs and node B detects it immediately. However, node A will not see the collision until the corrupted frame reaches it at a time about t0 + 2a.

TOPIC 2

NETWORK ARCHITECTURE

33

t0

A A begins transmission

t0 + a e

A B begins transmission

t0 + a

A B detects collision

t0 + a e

A A detects collision just before end of transmission

Figure 2.5: Collision in Ethernet [Source: Stallings (1991) Data and Data Communications, 3rd edn, New York: Macmillan College Publishing, p 398.]

If the transmission time for node A is shorter than 2a, node A will complete the transmission before knowing about the collision. This results in an unreliable transmission: Node A thinks that the transmission is successful but actually it is not. So, the transmission time has to be greater than 2a. This imposes a minimum frame size in the Ethernet. The following shows that the minimum size for an Ethernet frame should be 2aC, where C is the speed of the link in bits per second. Assume the time for transmit a frame be t and the frame size be s:

t = s/C
As discussed above, the transmission time has to be greater than 2a:

t >= 2a s/C >= 2a s >= 2aC

Example For a link of speed 10 Mbps and propagation delay of 51.2 10-6s: Minimum frame size = 2 51.2 10-6 10 106 = 1,024 bits.

34

TOPIC 2

NETWORK ARCHITECTURE

(a)

Ethernet Frame Types In the following reading from Dean, you are introduced to four kinds of Ethernet data frame: Ethernet 802.2 Ethernet 802.3 Ethernet II Ethernet SNAP

Each frame type differs slightly in the way it codes and decodes packets of data. On contemporary LANs, you can only focus on Ethernet 802.2 or Ethernet 802.3 frame types. As a professional network designer, you will need to improve a networks performance by identifying the kinds of frame that traverse the network. Sometimes, the most common problem relating to frame types arises from incompatibility between what a workstation expects to receive and what the server transmits.

Reading
Dean, pp. 2039.

(b)

Ethernet Design Consideration Ethernet is usually implemented with coaxial cable. Ethernet uses a bus topology, so any node can be easily tapped into the bus. On coaxial cable, a T-connector is used to branch out to a node. Each node should be at least 2.5 metres apart; otherwise, signal reflection may give a false collision indication. Each cable segment should not exceed 200 metres. If you want to build a network in an area exceeding this limit, you may use a repeater to extend the network. A repeater will receive signals from one cable segment, amplify and transmit it to another, as shown in Figure 2.6. However, you may not use more than four repeaters between any two nodes in the network. As introduced in the last section, a signal sent out from a node will propagate in both directions along the bus until the end. It is the same in Ethernet implementation. A terminator at the end absorbs the signals and thus prevents it from bouncing back and interfering with the trailing signal. Nowadays, UTPs are also commonly used in Ethernet.

TOPIC 2

NETWORK ARCHITECTURE

35

Repeater Weakened signal Regenerated signal

Figure 2.6: Repeater amplifies attenuated signal [Source: Microsoft Corporation (1996) Networking Essentials, 2nd edn, Microsoft Corporation.

Ethernet encodes the bits into electronic signals. The transmission rate for Ethernet is 10 Mbps but this number does not reveal the full facts. If only one user is using the network, then a 10-Mbps transmission rate can be achieved. However, if two users are using the network, collisions may occur and retransmission is needed, so the effective transmission rate drops. If 10 users are on the network, the effective transmission rate is approximately 7 Mbps. This drops again to 3.5 Mbps for 100 users. Newly developed Ethernet technologies have a faster speed. Fast Ethernet supports 100 Mbps and is now emerging as a standard. UTP Category 5 cable is used to implement the network. Developed by an alliance of 28 companies, Gigabit Ethernet supports 1,000 Mbps. UTP Category 5 cable is also used in its implementation with a length limit shortened to around 25 metres.

2.2.5

Token Ring

Token Ring is another network technology used for sharing a common link on a network. It has many different implementations such as IBMs 16 Mbps Token Ring, IEEE 802.5 and Token Ring. This section illustrates the data flow in a Token Ring network. Token Ring is implemented on a ring topology, in contrast to the bus topology in Ethernet. For a ring topology, nodes are connected with several point-to-point links, and frames flow in one direction a node receives frames from its upstream neighbour and passes them to its downstream neighbour, and copies the frame to its internal memory if it finds the frame is destined to itself.

36

TOPIC 2

NETWORK ARCHITECTURE

What algorithm does Token Ring use in controlling which node sends data first? It uses a token. A token is actually a special sequence of bits circulating around the ring as ordinary frames. Any node that wants to send data out must hold the token. In other words, nodes on a Token Ring network transmit data in turns. When a node has data to send and the token arrives at it, it takes off the token (that is, does not pass it to its downstream neighbour) and sends the data out. The data is then passed to the nodes on the network until reaching the destination node which copies the frame of data into its internal memory. The frame continues to travel after reaching the destination node and until it comes back to the sender node which absorbs the frame and releases the token back to the network. You may refer to Figure 2.7 for a view of this operation.
B

Free token

Sender A looks for token. When it comes around, A captures it.

D B

A Data D B Data A

A sends frame and it is received by C.

Free token

When first bit of the frame loops back, A regenerates token. A absorbs its transmitted frame.

Figure 2.7: Data flow in a Token Ring network

TOPIC 2

NETWORK ARCHITECTURE

37

The next reading describes the operation and architecture of a Token Ring network. Although Ethernet is much more popular than Token Ring in network implementation mainly due to its low costs and simple configurations, it is still an important network transport model.

Reading
Dean, pp. 21116.

Reading 2 (optional)
You will find a graphical view of how a Token Ring network works on the following homepage: http://www.rad.com/networks/1997/nettut/token_ring.html

ACTIVITY 2.1
You have been contacted by some representatives of the Malaysian Government for help in making a decision. They want to network government offices in the 10 largest districts in Malaysia in order to provide services to the communities. The offices have 50 to 250 workstations each and need to connect to one another to provide critical services. Their question for you is whether to use Ethernet or Token Ring communications. Make a list of the advantages and the disadvantages of each. Then, in your next tutorial, make a recommendation to your tutor, based on your list.

SELF-TEST 2.2
1. 2. 3. 4. Why is packet switching more efficient than circuit switching? What are the two potential causes of collisions in Ethernet networks? Compare and contrast CSMA/CD and token passing as access methodologies. What is the difference between a logical and a physical topology?

38

TOPIC 2

NETWORK ARCHITECTURE

2.3

NETWORKING HARDWARE

In data networks, hubs, routers, bridges and switches handle the task of directing information to the correct destination. In this section, you learn about these devices and their roles in managing data traffic. While Topic 1 focused on the physical layer of the OSI model, this section delves into the data link and network layers. It introduces the concepts involved in moving data from place to place by using different network devices such as repeater, hub, bridge, switch, router and gateway. In addition, you will learn all network interface cards: the workstations link to the network.

2.3.1

Network Interface Cards

Network Interface Cards (NICs) are interfacing hardware for computers to connect with a link. They are responsible for converting computer data to a form that can be transmitted over the links, and converting incoming data from the link to a form that can be understood by the computer. NICs come in a variety of types, depending on the network transport system (Ethernet versus Token Ring as discussed in previous section), connector interfaces (e.g. BNC versus RJ-45), type of compatible system board and manufacturer. Your next reading describes four primary bus architecture standards: Industry Standard Architecture (ISA) Extended Industry Standard Architecture (EISA) Micro Channel Architecture (MCA) Peripheral Component Interconnect (PCI)

You are shown how to install and configure a NIC. Several factors that should be considered when choosing an NIC for your workstation will also be considered. The most critical factor is compatibility with your existing computer system.

Reading
Dean, pp. 24062.

TOPIC 2

NETWORK ARCHITECTURE

39

2.3.2

Hubs and Repeaters

A hub is the most primitive network device and is a good substitution for the conventional Bayonet Navy Connector (BNC) connection in network configurations. In order to simulate a star topology, a hub concentrates the wires to a common point in the network. The hub allocates a port to each host and the hosts are interconnected through the hub. The hub works as a centre of distribution in matters of signals from and to the different parts of the network.
File Hub E-mail

Figure 2.7: Hub devices network configuration

As its name implies, a repeater copies or repeats signals that it receives. However, a repeater also amplifies all received signals before retransmission (without changing the frequency). The signal amplification occurs in the physical layer and the analogue signal (including noise and real data) will be amplified by the repeater. This happens in order to compensate for the attenuation that the waveform experienced during the journey to the repeater, before sending the signal on its way. The advantage of using a repeater is to extend the geographical distance of the network coverage. By strategically placing repeaters along a network bus, the distance between adjacent computers can be extended to connect a network over a larger area. Long-haul networks may contain many repeaters. Ethernet networks frequently use repeaters in order to extend the length of the bus within a LAN, meaning that the repeaters are connecting two segments of the same network to avoid attenuation problems. However, repeaters cannot be used to extend the network distance to infinity. The amplification of noise signals in repeaters will degrade the data quality. In order to obtain data of an acceptable quality, the number of repeaters connected to a network segment is limited to a specific number.

40

TOPIC 2

NETWORK ARCHITECTURE

In the reading that follows, you will see descriptions and illustrations of the operations of repeaters and hubs. With pictures, you can better understand the physical connection of those network devices to the network. Finally, the reading suggests several factors that you should consider in choosing the network devices.

Reading
Dean, pp. 26372.

Reading 2.1 (optional)

Description of repeaters http://www.rad.com/networks/1997/nettut/repeaters.html Description of hubs http://www.rad.com/networks/1997/nettut/hub.html

2.3.3

Bridges

Bridging technology is able to provide more features than hubs or repeaters. We can use bridges to develop a more complicated network. A bridge is a device that allows two networks that use the same technology to be connected. Bridging occurs at the data link layer, which controls data flow, handles transmission errors, provides physical (as opposed to logical) addressing and manages access to the physical medium. Bridges provide these functions by using various link layer protocols that dictate specific flow control, error handling, addressing and media access algorithms. Bridges examine incoming frames, make forwarding decisions based on information contained in the frames and forward the frames to the destination. Hence, a bridge improves the performance of a network system. As we know, a collision detection system requires a computer to delay transmission after a data collision occurs. The more computers connected to the network, the more collisions occur and the more delays suffered. The operations of bridges are transparent to upper layer protocols (network layer or above) and are not required to examine upper layer information. This means that they can rapidly forward traffic representing any network layer protocol. Because link layer information often includes a reference to an upper layer protocol, bridges can usually filter on this parameter. Furthermore, filtres can be

TOPIC 2

NETWORK ARCHITECTURE

41

helpful in dealing with unnecessary broadcasts and multicast packets. By dividing large networks into self-contained units, bridges provide several advantages, including the following: Because only a certain percentage of traffic is forwarded, the bridge diminishes the traffic experienced by devices on all connected segments. The bridge acts as a firewall for some potentially damaging network errors. Bridges allow for communication between a larger number of devices than would be supported on any single LAN connected to the bridge. Bridges extend the effective length of a LAN, permitting the attachment of distant stations that were not previously connected. Transparent Bridging This is another type of bridging technology. Transparent bridges are so named because their presence and operation are transparent to the network hosts. Transparent bridges are able to learn the networks topology by storing the source address of incoming frames from all attached networks. If, for example, a bridge sees a frame arrive on line 1 from host A, the bridge concludes that host A can be reached through the network connected to line 1. Through this process, transparent bridges build a table such as the one in Figure 2.8.
Host address A B C D E F G Network number 1 1 2 2 1 1 3

Figure 2.8: Example of a transparent bridging table

The transparent bridging table is used for traffic forwarding. When a frame is received on one of the interfaces of the bridge, the bridge looks up the frames destination address in its internal table. If the table contains an association between the destination address and any of the bridges ports other than the one on which the frame was received, the frame is forwarded out of the indicated port. If no association is found, the frame is flooded to all ports except the inbound port. Broadcasts and multicasts are also flooded in this way.

42

TOPIC 2

NETWORK ARCHITECTURE

By using the transparent bridging technology, a LAN can be separated into different segments. Instead of traversing incoming frames to all other interfaces of the bridge, the reduction of unnecessary traffic forwarding can improve network response times as seen by the user. The extent to which traffic is reduced and response times are improved depends on the volume of intersegment traffic relative to the total traffic, and the volume of broadcast and multicast traffic. The next textbook reading describes the operations of bridges in general. You learn how bridges manage data traffic and direct information to the correct destination.

Reading
Dean, pp. 27375.

2.3.4
(a)

Switches

Switching Basics The purpose of switching is to be able to reduce traffic congestion in different network topologies like Ethernet and Token Ring. To improve network response time, we have to reduce broadcast traffic and increase bandwidth. A switch can achieve this purpose by filtering broadcast traffic, whereas switching technology is able to increase the bandwidth of the network. In order to reduce the traffic flood between different segments, switches apply technologies like bridges. Switches store a table of physical addresses to determine the segment on which a datagram needs to be transmitted, thus reducing traffic. Since the operating speed of switches is much faster than that of bridges, the performance of switching technology is better able to support high-volume traffic on networks. Switches differ in the method of switching they use cut-through mode or store and forward mode. Cut-through mode: A switch running cut-through mode will read a frames header and decide where to forward the data before it receives the entire packet. Store and forward mode: In store and forward mode, a switch reads the entire data frame into its memory and checks it for accuracy before transmitting the information.

TOPIC 2

NETWORK ARCHITECTURE

43

(b)

Higher-layer Switches You learned that switches operate in Layer 2 (data link layer) of the OSI model, routers operate in Layer 3, and hubs operate in Layer 1. As switch technology advances, manufacturers are already producing switches that can operate at Layer 3 (network layer) and Layer 4 (transport layer), making them act more like routers. A switch capable of interpreting Layer 3 data is called a Layer 3 switch. Similarly, a switch capable of interpreting Layer 4 data is called a Layer 4 switch. These high-layer switches may also be called routing switches or application switches. The ability to interpret high-layer data enables switches to perform advanced filtering, statistics keeping and security functions. Layer 3 and Layer 4 switches may also transmit data more rapidly than a router, and may be easier to install and configure than routers. However, these switches arent as fully featured as routers. For example, they typically cannot translate between Token Ring and Ethernet networks, encapsulate protocols, or prioritise traffic. These critical differences make switches inappropriate for specific connectivity needs. The following reading describes the operation, advantages and disadvantages of switches. Two methods of switching cut-through mode and store-and-forward mode are discussed. High-layer switches are yet another technology you will need to explore closely.

Reading
Dean, pp. 27580.

2.3.5
(i)

Routers

Router Features and Functions A router is a device which routes data between networks of different technologies such as Ethernet, Novell or IBM Token Ring. Special features of routers are that they: Provide support for multi-protocols; Provide multi-interfaces for different networks; and Operate at the Internet protocol layer.

44

TOPIC 2

NETWORK ARCHITECTURE

Routers consist of multiple interfaces to transfer information across an internetwork from source to destination. However, the Internet is not a network with homogeneous technologies. It is necessary to build some network devices that are able to interchange data between different technologies. Routers are designed for this purpose. A router has an address on the network; a bridge does not. Networks frequently use a router as an intermediate destination. There are thousands of routers linking the world through the Internet. A router examines packets of data if they are addressed to it. After checking the packets destination protocol address, the router determines whether it knows or does not know how to forward the packet to the next hop. If the router does not know how to forward the packet, it typically drops the packet. If it knows how to forward the packet, it changes the destination physical address to that of the next hop and transmits the packet. If the next hop is not the final destination, a similar process will continue in the following hop until it reaches the final destination. (ii) Bridges versus Routers The major differences between bridges and routers are that: A bridge has to examine all data in order to determine the destination. If a source and destination are in the same network segment but on different interfaces of a bridge, data from the source have to pass through the bridge in order to reach the destination. By doing so, it could create a bottleneck if the capability of the bridge is not sufficient. However, the capabilities of routers are more powerful than those of bridges. Moreover, routers work at a higher layer. Instead of checking the physical address at the data link layer, as bridges do, routers examine the network address of each incoming data block in order to determine the corresponding path. Since routers are able to provide network layer services, data from different network segments can be exchanged via routers. Thus, routing and bridging accomplish their tasks in different ways. The reason is that the capabilities of routers are more powerful than those of bridges. Bridging occurs at the data link layer of the OSI reference model, whereas routing occurs at the network layer. All interfaces in routers have corresponding addresses. It is possible to connect to routers through their interface addresses.

(iii) Routing Components Routers are intermediate devices that perform the routing operation. Routing involves two basic operations:

TOPIC 2

NETWORK ARCHITECTURE

45

Determination of optimal routing paths; and Transportation of information groups (typically called packets) through an internetwork.

The cost of the routing operation is weighed by a number of factors, for example the path length that is used by routing algorithms to determine the optimal path to a destination. Routing algorithms generate and maintain routing tables to search the optimum path. The routing tables contain route information that varies depending on the routing algorithm used. A routing table stores the information for the next hop of the network. Destination or next hop associations tell a router that a particular destination can be reached by sending the packet to a particular router representing the next hop on the way to the final destination. When a router receives an incoming packet, it checks the destination address and tries to associate this address with the next hop. Figure 2.9 shows an example of a destination / next hop routing table.
To reach network 27 57 17 24 53 16 36 Send to Node A Node B Node C Node A Node A Node B Node A

Figure 2.9: Destination/next hop routing table

Besides saving the next hop information, routing tables contain information about the cost for each path. If multiple paths are available to reach the destination, routers compare metrics to determine the optimal routes for the packets. Metrics differ depending on factors such as interface speeds, reliability and the path length. In order to perform these routing functions, routers must talk to one another (and maintain their routing tables) through the transmission of a variety of messages. The information stored in the routing table is constantly updated. Routers compute the routing path based on the information stored in routing tables. Path length is the most common routing metric. Some routing protocols allow network administrators to assign arbitrary costs to each network link. Here, path length is the sum of the costs associated with each link traversed. Other routing protocols define

46

TOPIC 2

NETWORK ARCHITECTURE

the hop count, a metric that specifies the number of passes through internetworking products that a packet must take on the route from a source to a destination. (iv) Brouters and Routing Switches The network industry has adopted the term bridge router or brouter to describe routers that take on some characteristics of bridges. The advantages of crossing a router with a bridge is that you can forward nonroutable protocol, such as NeBEUI (the protocol is further discussed in Unit 3), plus connect multiple network types through one device. A bridge router offers support at both Layers 2 and 3 of the OSI model. It intelligently handles any packets that contain Layer 3 addressing information and simply forwards the rest. Another router hybrid, a routing switch, combines a router and a switch. It can interpret data from both Layers 2 and 3 of the OSI model. A routing switch is not as fully featured as a true router, and therefore routing switches have not gained wide acceptance from networking professionals.

2.3.6

Gateways

Gateways do not neatly fall into the networking hardware category. They are combinations of networking hardware and software that connect two dissimilar kinds of network. Specifically, they may connect two systems that use different formatting, communications protocols or architecture. They repackage information so that the data can be read by another system. Gateways must operate at multiple layers of the OSI model. They must communicate with an application, establish and manage session, translate encoded data, and interpret logical and physical addressing data.

ACTIVITY 2.2
Please use the Internet to view Cisco Systems new product release bulletins at http://www.cisco.com/warp/public/cc/general/newprod.htm. What networking trends do you notice from reading these bulletins? Discuss your findings with your tutor during the next tutorial, or email another student and compare your findings.

TOPIC 2

NETWORK ARCHITECTURE

47

SELF-TEST 2.3
1. Assume your office is running an Ethernet LAN. If you have to connect five additional stations on the next floor into the existing network, what is the most effective way to do so? State the major differences between bridges and routers. What happens if the convergence process of routers is not fast enough?

2. 3.

Activity 2.3: Case study

Acme Accountants Ltd the networked office
Scenario Acme Accountants Limited moved to a new office in Central just one month ago. The office occupies a single floor of about 5,000 square feet, which accommodates 100 employees and will serve as the headquarters of the company. At the same time, they want to implement an office automation project. Phase I is to network their 20 existing standalone computers, so that they can use a newly distributed accounting software application. The software will be installed on a dedicated server machine, and staff will access features of the application from one of the 20 existing computers. The budget for this project is small but adequate. Acme needs to determine its basic networking requirements so that the work can be put out to tender. Questions 1. Cabling Assuming Acmes computers have not been networked before, what additional hardware might they need to have fitted? And what kinds of link (digital or analogue) do they need to connect the computers? Network topology Acmes new distributed software application will require multiple user access to the server at any one time. The seating plan for the staff has not been finalised, so the relocation of computers has not yet taken place. Furthermore, the company has planned that 30 new computers will be bought and networked to serve more staff in phase II of the project, on the

2.

48

TOPIC 2

NETWORK ARCHITECTURE

successful completion of phase I. What type of network topology would you recommend for Acme? Since network topology is closely related to the cabling and the network technology used, you may have different advice after studying the next section and the section on network technologies. 3. Network adaptor All of Acmes existing computers are Pentium PCs. What type of network adaptor would you recommend they buy, and why? Network adaptors should be matched with the network technology used, so you may specify which network technology the network adaptor is for after studying the section on network technologies further on in the unit. The PCs are spread around the office. The office already has a raised floor for cabling to be fitted underneath. As well as supporting more users, three more shared applications on personnel management, financial analysis and inventory will be installed on the server in phase II. What type of cabling would you recommend for Acme, and why would you recommend it? 4. Network technology In phase III, the final phase of the office automation project, all staff at Acme will be provided with a networked Pentium PC. It is also expected that there will be 10 applications running on the dedicated server for shared use. One more server will be installed for back-up. There may also be heavy sharing of data among the users PCs. What would you recommend for Acme: Ethernet or Token Ring? Why would you recommend this option?