9.1 Network Topologies

Various arrangements of computers and cabling can be used to accomplish the task of enabling communication to take place between computers. These arrangements are called network topologies. Each arrangement has advantages and disadvantages. In this section, different topologies are discussed in detail to understand its pros and cons. The different topologies allow the computers in a network to communicate without individual point to point connections, but with reduced number of connections between the various computers.  To make this happen, the information flow over the physical connections need to be controlled carefully. Devices that perform this control of the information flow from one computer to another is called a network device. In a network there are two types of devices: network devices and terminal devices. Network devices are devices that are used in the running of the network, while terminal devices are devices that use the service provided by the network. This is similar to an electrical power network where some devices like transformers and switches are devices that are used in the running of the network, while a light bulb, computer, and television are terminal devices, essentially users of the power supplied by the network.

Point to Point Connections

When two computers are connected directly to each other, then it is called a point to point connection. This connection is a dedicated connection that allows information to flow between the two computers at any time, as long as the connection is established.

A point to point connection between two computers could be considered the simplest case of a network because it facilitates information transfer. In this case a point to point connection could be established by providing a connecting wire between the computers to carry the information from one to another. To achieve this, hardware and software are needed on both computers to establish the communication. This type of connection could be established using the serial or parallel ports on the computer. If the computers are physically close to each other, then the task is easy because wiring could be easily done to provide point to point communication.

On the other hand, if the computers are located far apart then the wiring from one to the other becomes a challenge. Imagine connecting two computers, one in Toronto and the other in New York. The cabling from Toronto to New York needs to be established to send information back and forth. Providing cabling for such a distance is costly. So, an alternate solution was created. During the 1960's, the telephone network was well established with cables going from one place to another; thus, the idea to use the existing telephone network to transfer information instead of laying new cables developed. This idea lead to the development of the MODEM (modulator - demodulator). The modem essentially took the data and converted it into a form that could be transmitted and received over the existing telephone lines.

When computers are connected using point to point connections, then a simple computer network is established. The disadvantage of this type of network is that as the number of computers in the network increases, so does the number of connections required. As evident in figure 9.1. This creates the problem of excessive wiring costs, and therefore becomes impractical when dealing with a large number of computers.

From the above point to point connection diagrams for different number of computers, it is evident that the number of point to point connection is going to increase with the number of computers in the network. The rate at which the number of connections increases is much faster than the rate of increase for the computers. For example, the number of point to point connections for a network with 20 computers will be 180 connections, and for 100 computers 4950 connections. The problem with this type of approach is that each point to point connection will increase the overall cost of the network. This is similar to the power supply situation discussed earlier. To solve this problem the network should be designed differently without using point to point connections.

If information transfer is to take place between computers, then a connection of some sort needs to be present to provide a path for the information flow. Imagine a simple network of four computers. In such a case the connection could be established in several different ways, as shown in figure 9.2. In all the different arrangements shown there exists a path between any two computers for the flow of information.

Though the connections are visually different, some of them are similar in nature. For example connection option B and F are similar in that they are both serial in nature, and connection option C and D (assuming no connection in the middle) are similar in that they both have a ring pattern. Connection option E and H are also similar in that each computer has only one connection to it. These similar connections are grouped together and called topologies because the information flow control requirements for each topology will be the same.

The Star Topology

In the star topology the devices are arranged in the form of a star as shown in figure 9.3. In this arrangement, the terminal devices are connected to a central a network device, called a hub, at the center. In this arrangement the hub is responsible for the flow of information from one computer to another, and all the traffic must pass through the hub in the middle.

The number of connections required for this arrangement is now 6 instead of the 15 connections that were needed for the point to point arrangement. Similarly if the number of computers were to increase to 100 then the number of connections would be 100, instead of the 4950 that is required for a point to point setup. This reduction in the number of connections will translate to a reduction in the cost of implementation.

In the star topology the network hub is responsible for controlling the flow of information from one terminal device to another. The controlling process has to be based on a set of rules that will facilitate the transfer of information from one device to another. The details of the control and transfer of information will be discussed later. Each terminal device on the network is capable of communicating with another through the central hub. The problem with this arrangement is that if the hub of the network fails then the whole network fails. This arrangement is prone to network failure due to its centralized structure. This type of topology was used in the ARCnet networks during the 1980s.

The Bus Topology

The bus topology is similar to a road with all the driveways of the houses connected to it. If one wants to travel from one house to another, then one would enter the roadway through the driveway and get off the roadway at the destination driveway. Here there is a common central line for the network traffic, and all the terminal devices are connected directly to it. The central line or traffic carrier is called the network bus.  In figure 9.4 the arrangement shows a bus topology

In this arrangement, communication between two devices can take place through the central line or the bus. In this case the control of the data flow is handled by the individual terminal device instead of a central hub. Control is important because if all the computers on the network are sending information though a common line, then there may be a mix up of the information during the communication. With this arrangement, the information is placed on the bus by the transmitting devices, and it is seen by all the other devices on the network. The transmitter sends the information with the recipient’s address, so the information is processed by the device that has the same address as the transmission. This type of topology is used by Ethernet networks, developed by Xerox in the 1980s.

The Ring Topology

In the ring topology, the devices are connected in the form of a ring. This appear similar to the bus topology, where the two ends of the bus are connected. Though there is a physical similarity, the data transfer is implemented differently. The ring arrangement uses the idea of token passing, where the information is sent to the adjoining computer which then passes the information to the next and so on until the information arrives at the destination computer. In this topology each computer waits for its turn to transmit information and receive information. This type of topology is used in IBM’s Token Ring Network, which uses the idea of token passing to transfer data.

The Mesh Topology

The point to point connection can be viewed as an ideal case of the mesh topology, but any network where there are one or more interconnections between devices is also called a mesh network.

In this arrangement, communication between two devices can take place through the different devices on the network. The control of the data flow is handled by the individual devices on the network. For example, if information needs to be sent from device C to device D, then it could be sent to B and then to D or it could be sent to E and then to D. In this case, only a path is needed from one device to the other to communicate. One of the early mesh networks was the ARAPNET of the US Defence Department, which later went on to become the internet. Mesh networks provide flexibility because computers could be connected to the network in any arrangement.

9.1 Practice Questions 1.     What are some of the advantages of using a computer network? 2.     Explain what is meant by the term “point to point connection”. 3.    What are the disadvantages of a point to point networking system? 4.     List 4 services that are provided by networks. 5.     What are the 3 main types of network topology? 6.     Describe how information is transferred in a star topology. 7.     What are the advantages of the bus topology? 8.    Can the arrangements shown below be categorized as the same topology? 9.     Compare the fault tolerance (the ability to tolerate failure) of the various network topologies.