Network topologies are the logical or physical structure / layout of a network. Different topologies have their own pros and cons so it is important to know the differences between them and what situations they work best in.
Logical network topologies
Logical topologies are not about how the devices on a network are physically linked, instead they are about how the different nodes on the network communicate and how their signals act. For instance, logical topologies such as bus topology and ring topology, show how the signals on the network behave but in actuality can be laid out in different ways so can have separate physical topologies.
In a logical bus topology the data can be transmitted by one machine to the whole network via the bus and each networked machine will check to see if the data is intended for that recipient or not.
In a logical ring topology a token is passed around in a sequence and only the machine that the token currently resides with can actually transmit data to the network. This is to avoid data collisions by restricting transmission to one machine at a time.
A physical topology refers to how the actual machines are positioned on the network along with how they are physically connected e.g. types of cables etc. There are many different types of physical topologies with their own pros and cons, a few of which can be seen below:
This is a very simple setup in which every single machine is linked to a central cable which in turn means every single machine has a direct link to each other. This is a quick and easy way to set up a simple peer to peer network. It is quick to set up and cheap as it requires the least amount of physical cabling out of all the other topologies. However, it is a difficult topology to expand as it the efficiency / performance of the network degrades with an increase in workstations connected to the central cable.
A ring topology is when all the networked computers are linked in a closed loop. Each workstation is directly connected to a computer on either side of it, then further connected indirectly to the other machines on the network through the loop. A ring topology is another one that can be set up fairly easily and provided the machines that need to be networked have been positioned relatively close to one another in an organized manner the cabling costs of this type of network will still be one of the cheapest. The further apart, and more scattered, all the linked machines are the more time consuming and costly this type of network will become.
This is another type of network commonly used in peer to peer setups.
Star topologies are commonly used for the client / server type of network. A star topology will always have one central computer, commonly this central computer is the main server. All the other networked machines will be linked to this central machine, branching out like the points of a star.
In a star topology adding another machine is quite simply a case of connecting another branch to the main central computer. Unlike other topologies, the performance of the network does not degrade as more machines are added.
While a star topology is more expensive and complex to setup than previously mentioned topologies, if there is a cable failure, only that end machine will be removed from the network, making it more reliable than other topologies.
A tree topology is a network of multiple star networks. In a tree topology two or more star topologies are connected by linking their main / central machines to a central bus. Whole star networks can be added to the tree simply by connecting the main / central machine to the bus. Adding any single machine to the network is simply a case of connecting it to its corresponding central machine, in the same way you would expand a single star network.
Tree topologies are expensive and complex, however they are great in situations that require multiple star networks to be connected. For instance, companies or schools that may have multiple buildings that each contain star networks within those individual buildings that need to be connected as a whole across the company site or school campus.
If a cable goes down to a single end machine, only that end machine will be effected and the rest of the network will continue to function. If a central machine connected to the bus fails, it has the potential for the whole start network within that particular building to fail but all the other star networks will continue to be networked and function as normally. If the main bus fails, it has the potential for complete network disruption.