1.2 Types of Network Topologies
Any computer network design can be said to have evolved from one of three basic topologies. These three topologies can also be combined in a variety of ways to form more complex hybrid topologies. These three topologies are:
A bus topology is commonly referred to as a “linear bus” because all of the nodes are physically connected in a straight line. A bus topology has a single backbone cable to which computers and other devices are connected. This backbone is also known as a segment or a trunk.
An Ethernet bus network, such as 10Base2, has a single communications channel with each node connected to the main cable. The bus must be terminated at either end by terminators. These absorb the signal and stop it being sent back along the cable to corrupt any messages. If the bus is not terminated the signal will reflect back and possibly collide with other data. When a node transmits its data it goes in both directions. If a node recognises that the message is addressed to it will take action, otherwise it will ignore the message.
Data Communication on the Bus
On a bus topology only one computer can send data on the bus at any one time. Therefore, the more computers added to the bus the more likely that collisions will take place and this will have an adverse effect on network performance, ie: the network will be slower.
All nodes on a bus have to contend with each other to be able to send messages and sometimes collisions will occur on this type of network. One way to deal with collisions is to use an access control protocol. One commonly used protocol is called Carrier Sense Multiple Access / Collision Detection (CSMA/CD).
Collisions are an unavoidable feature of contention based networks and can become a problem if too many stations are added. A happy medium is required where the line is not idle nor is it constantly experiencing collisions.
Computers on a bus network transmit data or listen for data but they are not responsible for moving data from one computer to the next. This means that if a computer fails the network will not be affected. However, if there is a fault with the backbone cable then the network will cease to function.
In the star topology, cables from each computer are connected to a central device known as a hub. Signals are transmitted from the sending computer through the hub to all computers on the network. This topology has its origins in the early days of computing when terminals were connected to a central mainframe computer.
The star network topology has the advantage of centralising resources and management, however, more cable is required than for other topologies. The star topology also has a central point of failure, that is, if the hub at the centre of the topology fails then the whole network will be down.
If one of the computers (or the cable that connects it to the hub) fails on a star network, only the failed computer will not be able to send or receive network data. The rest of the network will continue to function normally.
This type of topology is the most commonly used. Most star topologies are Ethernet networks, which mean that they use CSMA/CD as a means of media access.
Hubs have generally been replaced in star topologies by switches, which have the advantage of creating multiple collision domains and effectively creating collision free networks.
The picture shown is a classical ring topology and consists of a physical ring and a logical ring. Each computer is linked to two others to form a closed loop. Each node in this ring will act as a repeater by regenerating and cleaning up a signal before passing it to the next node.
In a ring network communication takes place by the use of a token which is passed from computer to computer around the ring and it is only the computer that has the token that can send data. A ring network is unidirectional, ie: data can only go in one direction.
Once a computer has the token it can send its message. Each successive computer on the ring will examine the message to check if they are the message’s destination. Once the message has been received the token is put back on the network for the next node that requires its use.
No matter what the size of the network or the distance covered there can only be one token on a ring network. Ring networks are known as deterministic because network access is determined by which workstation has the token.
The failure of one computer can have an impact on the entire network.