Does NothingDoes NothingDoes NothingDoes Nothing NowtAdminCase Studiesmodule1Module2Module3Module4Module5Module6Glossary  
nowtPolicy / StrategySoftware evaluationDatabase managementDistributed SystemsNetwork security and accountingHuman Computer InteractionHuman Computer InterfaceSoftware DevelopmentSoftware reliabilityPortability of Datanowt
NowtNetworksNetwork TopologyBridges and repeatersSecurity and accountingQuestionsNowt

Network Topologies

The network cable links nodes. A node is any element connected to the network that plays some part in making the system work. A file server, a printer server, the interface on a user PC are all network nodes. The standard network topologies are:-

Bus Network

Here nodes are attached to a central bus. in an arrangement that is used for LAN networks. There is the definite possibility that two nodes will attempt to transmit at the same time - this is called a collision. In an Ethernet network this is handled as follows. A node needing to transmit examines traffic on the bus and waits until there is none. It then pauses - so that there is a minimum space between frames. It then puts its data on the bus. If a collision occurs because another node starts to transmit at the same time then both nodes cease transmitting, wait for a random time interval and then start the process again.

Bus Network

Note that there are terminators at either end of the cable to soak up the signal - preventing reflection back down the cable. Advantages are that less cable is needed than token ring, network can be easily extended and that if one section of cable is faulty the network can be shortened and the remaining section used while the fault is repaired. Disadvantage is that behaviour of network under heavy traffic is unpredictable and long delays may occur. A frame will consist of identification of source and destination, data length (variable) and data plus error detection data.

Star Network

Star Network

This was once a common configurations when a central computer was used to link dumb or semi-dumb terminals. It is not often used now but could be used for LAN or WAN networks. It has the advantage that if one cable or node (other than the central node) is faulty then the rest of the network can continue to function.

Ring Network

Here nodes are linked in a ring. Data is passed from one node to the next. This topology is used for LAN networks. The most important ring network implementation is the IBM token ring. A token is passed from node to node. A station wanting to transmit data to another station must wait until its interface node receives the token. The data is attached to the token together with the destination and transmitting stations addresses. The token is then passed onto the next node.

Ring network

Note that the interface is shorted out if the network station is powered down (a shorting relay is held open by station's power supply so that removing power closes the relay shorting out the node). When the packet of information is received at the correct destination then the data is extracted and flags are set within the packet to indicate that the data has been correctly received.

The token then continues round the ring until it returns to the originating station. The token is then freed to carry other data. Monitor node responsible for soaking up token that is set to circulate endlessly because initiating or receiving station is no longer present. This system's performance levels can be determined for different traffic levels - it is deterministic - which is an advantage. It also has the advantage that since data flows one way only, this configuration is able to transmit large volumes of data between PC's or to handle communication between PC's and a larger computer. A disadvantage is cabling cost to complete the ring and difficulty in extending an existing ring

A token ring network may in fact look like a star network if a device called a hub is used. The hub contains the ring, including the relays that connect the computer to the ring. Cable is then run from each individual computer to the central hub. This overcomes some of the cabling costs associated with a ring.

This arrangement can be used in conjunction with a bus network. The bus network is used for the principal network that runs through the building - which could be a high capacity fibre optic cable. This principal network is called the backbone. At particular points on the network a hub may be attached to provide access for a number of computers in a particular work area.

Mesh Network

Mesh network nodes are connected one to another as required. This topology is used for WAN networks and data is usually transmitted in packets in what is known as a packet switching network.

Mseh network

The complete message to be sent will be broken up into a sequence of packets each containing source and destination addresses and a sequence number. These packets will be transmitted from node to node by whatever routes are free. They will probably arrive at their destination out of sequence because of the different routes taken. The receiving node will reassemble the packets into order before extracting the data. Note that in a message switching network a complete message is transmitted in one go. This requires a sender to destination link to be established before the transmission can begin. Packet switching offers a more flexible approach and allows the traffic to be more evenly spread over the network.

 

©LEV