Contents

1 Introduction 2

1.1 Ethernet Cables . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2 Repeater . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 1.3 Hub . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 1.4 Bridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 1.5 topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 2 Design 4

2.1 Thin Coax Implementation & Justification . . . . . . . 5 2.1.1 Thin Coax Implementation Cost . . . . . . . . . . . . . 6 2.2 UTP Implementation & Justification . . . . . . . . . . . .6 2.2.1 UTP Implementation Cost . . . . . . . . . . . . . . . . . 8 3 Summary and Conclusions 4 References 8 8

List of Figures
1.1 Ethernet Cables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Topologies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1 Bus Topology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Star Topology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Campus Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Thin Coax LAN Implementation . . . . . . . . . . . . . . . . . . . . 2.4 UTP LAN Implementation . . . . . . . . . . . . . . . . . . . . . . . .

Abstract
Throughout history economies have had certain areas of focus. In the agricultural age, whoever had the most food was the most powerful. In the industrial age, the nations with the most advanced manufacturing were the most successful. Sometimes contemporary times are referred to as the Information Age because the main resource that businesses rely on is the exchange of information using computer networks. It is therefore critical that all businesses have an economic and efficient computer network to be successful. The objective of this project is to design a LAN connecting four departments in a campus layout. Two designs meeting specific requirements are discussed in this report, one using thin coax and the other using UTP. Arguments justifying the design and correctness, together with the total implementation costs are presented for each implementation. Fiber optic cable is used in the network backbone because it offers high data rate capacity, no electrical interference, ability to withstand severe weather conditions and larger maximum segment length.

1 Introduction
Computer networks are essential for day-to-day communication, e-commerce transactions, information sharing and the like. Networks can be broadly categorized according to their size, which helps determine their scope and intended application. Different networks are implemented in different ways. Network: It is a set of devices(nodes) connected by a communication link. Criteria that the network must meet are following: 1) Performance: it is the parameter that is evaluated by 2 networking metrics i.e. throughput that should be more and delay that should be less. 2) Reliability: time taken by a link to recover from a failure. 3) Security: it means to prevent the unauthorized access of the network. Local Area Network (LAN) is a high speed computer network used for connecting computers and sharing resources over a relatively small area (few kilometers). Typically, LAN can be implemented in an area as small as a dorm room to one that is as large as a university campus. In this project, we will design a wired LAN for one such campus layout. We will use two different cabling technologies and topologies for our wired LAN. The First: Ethernet design will be implemented using thin coax cable on a bus topology in which all devices will be connected to the central bus. The second design: UTP cable will be implemented using star topology in which all devices will communicate on the network by passing data through a hub.

1.1 Ethernet Cables

Unshielded Twisted Pair (UTP) cable consists of eight insulated copper wires grouped together to form four pairs. The wires in each pair are twisted around each other to reduce signal degradation from interference. The number of twists per meter for each pair can be varied to further reduce crosstalk noise. The Category 5 UTP (also 100BaseT) used for fast Ethernet, supports data rates of 100 Mbps or higher and has a maximum segment length of 100-meters. The standard connector for UTP is an RJ-45 connector. By far, UTP is the most popular and generally the best option for wired LANs. Thin Coax cable consists of an insulated copper wire surrounded by a woven copper braid to reduce the amount of outside interference. An outer jacket covers the two conducting elements.

Coax cable supports data rates from 10 to 100 Mbps and can be laid over greater distances than UTP with fewer repeaters. The thinnet cable used for Ethernet has a maximum segment length of 185-meters. Network interface cards (NIC) with the standard connector for coax cables, the British Naval Connector (BNC), are a rare find nowadays.

Fiber Optic cable consists of a center glass core surrounded by layers of protective plastic shield, followed by a very strong plastic Kevlar (used in making armor) and an outer jacket. The electrical interference is absent in fiber optics since it transmits light and not electrical signals. 100BaseF is the fiber optic specification used for carrying Ethernet signals. It supports data rates up to 10 Gbps and has a maximum segment length of 2000- meters. The most common connector for 100BaseF is the SC connector, which has two optical connections, one each for connecting to transmit and receive.

(a) UTP Cable and RJ-45 Connector

(b) Coax Cable and BNC Connector

(c) Fiber Optic Cable and SC Connector

Figure 1.1: Ethernet Cables

The exceptional qualities of 100BaseF such as high data rate capacity, no electrical interference, ability to withstand severe weather conditions and larger maximum segment length, makes it an ideal choice for connecting networks between buildings. Figure 1.1 illustrates all the three types of Ethernet cables discussed in this section.

1.2 Repeater
A repeater is a Physical Layer (layer 1) analog device that connects two cable segments. It operates on bits on the cable segments, amplifying their signal strength before sending it to the other segments. Classic Ethernet allowed four repeaters to join five segments together. Repeater has no filtering capability due to which it forwards every frame. The concept is still in use, mainly for reliable data transfer.

1.3 Hub
A hub is also a Physical Layer (layer 1) device and somewhat similar to the repeater; it operates on bits and the number of hubs is limited to four for a joining segments together. It is mainly used for UTP networks to provide connection between nodes. A hub does not isolate traffic because two signals arriving at the same time on a hub will collide and will have to be retransmitted. Hubs form a single collision domain.

1.4 Bridge
A bridge is a Data Link Layer (layer 2) device and unlike repeaters and hubs, it operates on frames of data. Bridges use look-up tables to route the data frames to the destination address after reading it from the frame header. Unlike hubs, bridges manage traffic to avoid collisions. As a physical layer (layer 1), it act as a regenerator. As a data link layer(layer 2), it can check the physical(MAC) addresses(source and destination) contained in a frame. It has the filtering capability. We have 2 types of bridges: 1) Transparent bridge: A transparent bridge is a bridge in which the stations are completely

unaware of the bridge's existence. If a bridge is added or deleted from the system,

reconfiguration of the stations is unnecessary. According to the IEEE 802.1 d specification, a system equipped with transparent bridges must meet three criteria: I. Frames must be forwarded from one station to another. 2. The forwarding table is automatically made by learning frame movements in the network. 3. Loops in the system must be prevented.
2) Source routing bridge: Another way to prevent loops in a system with redundant bridges is to use source routing bridges. A transparent bridge's duties include filtering frames, forwarding, and blocking.

In source routing, a sending station defines the bridges that the frame must visit. The addresses of these bridges are included in the frame. In other words, the frame Contains not only the source and destination addresses, but also the addresses of all bridges to be visited. The source gets these bridge addresses through the exchange of special frames with the destination prior to sending the data frame. Source routing bridges were designed by IEEE to be used with Token Ring LANs. These LANs are not very common today.
1.5 Topologies

A backbone network allows several LANs to be connected. In a backbone network, no station is directly connected to the backbone; the stations are part of a LAN, and the backbone connects the LANs. The backbone is itself a LAN that uses a LAN protocol such as Ethernet; each connection to the backbone is itself another LAN. Although many different architectures can be used for a backbone, we discuss only the two most common: the bus and the star.

Bus Backbone: In a bus backbone, the topology of the backbone is a bus. The backbone itself can use one of the protocols that support a bus topology such as lOBase5 or lOBase2. In a bus backbone, the topology of the backbone is a bus. Bus backbones are normally used as a distribution backbone to connect different buildings in an organization. Each building can comprise either a single LAN or another backbone (normally a star backbone). A good example of a bus backbone is one that connects single- or multiple-floor buildings on a campus. Each single-floor building usually has a single LAN. Each multiple-floor building has a backbone (usually a star) that connects each LAN on a floor. A bus backbone can interconnect these LANs and backbones. Figure 1.2 shows an example of a bridge-based backbone with two LANs.

Figure 1.2.1

Star Backbone: In a star backbone, sometimes called a collapsed or switched backbone, the topology of the backbone is a star. In this configuration, the backbone is just one switch (that is why it is called, erroneously, a collapsed backbone) that connects the LANs.
In a star backbone, the topology of the backbone is a star; the backbone is just one switch.

Figure 1.2.1 shows a star backbone. Note that, in this configuration, the switch does the job of the backbone and at the same time connects the LANs.

Figure 1.2.2

Star backbones are mostly used as a distribution backbone inside a building. In a multifloor building, we usually find one LAN that serves each particular floor. A star backbone connects these LANs. The backbone network, which is just a switch, can be installed in the basement or the first floor, and separate cables can run from the switch to each LAN. If the individual LANs have a physical star topology, either the hubs (or switches) can be installed in a closet on the corresponding floor, or all can be installed close to the switch. We often find a rack or chassis in the basement where the backbone switch and all hubs or switches are installed.

Marketing 15 hosts

Administration 30 hosts

WC 1

WC 2

Library 72 hosts
WC 3

WC 4

Engineering 15 hosts

Figure 2.2: Campus Layout

DESIGN REQUIREMENTS:

The project has the following design requirements: Each department must have access to the resources of all other departments. The traffic generated by the users of one department cannot affect another departments LAN unless accessing a resource on that other departments LAN. A file server can support only 30 users. File servers may not be shared by multiple departments. All repeaters, bridges, and hubs must reside in the wiring closets (WCs).

Marketing 15 hosts

Administration 30 hosts

Library 72 hosts

Engineering 15 hosts

Thin Coax Design Description and Justification:

Our implementation for the campus layout using thin coax is shown in Figure 2.3. The hosts are connected on a bus topology with the thin coax connected via the T-connector in a daisy-chain fashion. All hosts numbered 15, 30, 24, 48 and 72 in Figure 2.3 are terminated with a 50 BNC terminator.

The design uses a fiber as the backbone between all wiring closets. In figures 2 and 3 above, fiber are the lines in red and thin coax lines are in blue. It is necessary to use fiber in the 200 meter conduit because thin coax has a maximum range of 185 meters, and because repeaters may not be placed within the conduit. While the distance between WC3 (Library) and WC1/WC2 (Administration/Marketing) does not reach 185 meters, having a fiber backbone makes the network more futureproof because of the potential need of fibers high bandwidth. The two port bridges act as a layer 2 device to separate the traffic between the

campus departments, but still allow all the departments to communicate to each other. In Administration (top right), the coax splits between the file server and a 2 port repeater because there is a maximum of 30 connections per coax cable, while there are 31 connections (30 hosts, 1 file server). In Library, the coax goes from the 2 port bridge to a Multiport Repeater (8 thin coax) again because of the limitations of thin coax.

Table 1 No. of NIC thin coax ports:

In order to connect all the hosts and the file server on the network, the number of required NIC/coax ports is # Hosts (15 + 15 + 30 + 72) + #Servers (6) = 138.

The Thin Coax distance: The approximation was based on the user having enough wire to place machines anywhere in the room. Because Coax can branch off with T connectors in sort of series, the approximate sides of the rooms were added with 3 meters added per user. For example, look at the Engineering section in Figure 3. The thin coax line exits WC4 and travels downward. This section was estimated to be 40 meters because while the wiring closet consumes space, there still needs to be wire between the bridge and the hosts. Then 40 meters for the bottom and the left side of the square were used. There were no computers allotted in the top section. The wire going to the last host was just for

graphical representation. The distance between the main thin coax line and each host/server was estimated to be 3 meters. Administration and Marketing follow a very similar calculation. The exception is that administration requires an additional 5 meters for the file server. For Library, there needs to be 100 meters for the bottom part that has 30 hosts. The section that has 3 file servers and 12 hosts needs roughly 30 m for the right side, and 50 m for the right half of the top side, totaling 80 meters. The line that has 30 hosts on the top left needs about 40 m to get to the first host, and then 50 meters for the top right, and 30 m for the left side, totaling 120 meters. Engineering: 40+40+40 + 3*16 = 168 m Administration: 50+30+30 + 3*31 + 5 = 208 m Marketing: 30+50+30 + 3*16 = 158 m Library: 100+80+120 + 3*75 = 525 m Total = 158+193+148+515 = 1059 m The Fiber Distance:
A fiber optic cable repeater is used as a backbone for the thin coax network. The backbone is used to provide network reliability and to make it robust. Four 2-port Fiber/Coax bridges are used to connect each department to the network.

The fiber distance is found by taking the distance on the side of the walls with additional distance for slack. Conduit: 220 meters (20 slack) WC2 to Marketing: 105 (5 slack) + 35 (5 slack) = 140 meters WC2 to Support: 35 (5 slack) + 55 (5 slack) = 90 meters Total = 450 meters

Marketing

Engineering

Administration

Library

Thin Coax LAN Implementation

UTP Design:

Marketing

Administration

Library

Engineering

UTP Design Description and Justification:

The UTP design in similar to the Thin Coax in that it relies on a fiber backbone. In figures 4 and 5 above, blue lines are UTP while red lines are fiber. The positioning of the hosts are not as explicit in this design because each host has its own physical connection to a hub, and so I estimated that the average UTP cable length is equal to the distance to the center of the room. Thus, the hosts are roughly uniformly distributed and so explicit positioning was not possible. The maximum distance of UTP is 100 meters, and so a fiber backbone is even more necessary. All fiber lines connect to 2 port bridges that convert to UTP. After the conversion all lines connect to a 36 max connection UTP hub. In the Manufacturing sector, there are 3 UTP hubs because 75 connections are needed for 3 file servers and 72 hosts.

The calculation for the fiber length is the same. However, the UTP cable length is much higher. This is because every user has a dedicated line to the hub, which must be in the wiring closet. I calculated the distance by finding the distance to the center of the room, and reasoned that the average length would be towards the center if the computers were

uniformly scattered throughout the room. Even if the computers are not, this scenario is future proof to most reorganization of hosts. There would only be insufficient UTP cable if the room had most of its hosts toward the opposite corner of the wiring closet. Library: sqrt(502 + 152) * (75 connections) = 3915 m Engineering: sqrt(202 + 202) * (16 connections) = 453 m Administration: sqrt(252 + 152) * (31 connections) = 904 m Marketing: sqrt(252 + 152) * (16 connections) = 467 m Total = 5739 m

Table 2

Marketing

Engineering

Administration

Library

UTP LAN Implementation