Mesh, Bus, and Ring topologies
The first types of topologies I would like to talk about are the mesh, bus, and ring topologies. As far as mesh topologies go, there are actually two types.
There’s the full mesh, and then there’s also the partial mesh. In the full mesh topology, all the devices are directly connected to all other devices. This provides full redundancy for the network. While each device being connected to all other devices provides great redundancy, it also increases the cost significantly for the network.
The reason the full mesh node is so expensive is because each computer in the network is required to have multiple NICs and cables.
In fact, it’s required to have just as many NICs and just as many cables as there are computers in the full mesh network. The most likely place you’ll find a full mesh configuration would be in a WAN environment. These are seldom, if ever, used in LAN environments. This diagram illustrates what the full mesh environment looks like. Because it’s normally a WAN technology, I’ve used the example of cities to illustrate the nodes. As you can see, each city in this full mesh is connected to each other city in the network.
What that means in this case is the node in that city has to have three NICs to connect to the other three cities that it is connected to. With the full mesh, we also have a partial mesh. In the partial mesh, all devices are directly connected to at least two other devices. This provides strong redundancy for the network but not the full redundancy that the full mesh provides.
Because it doesn’t require the full redundancy, or it doesn’t require that every device is connected to every other device in the network but only a minimum of two, it’s not as expensive as the full mesh.
However, it is still a very expensive configuration to have. Again, just like with a full mesh, multiple NICs and cables are required for each machine connected into the partial mesh. Again, a partial mesh is most likely found in a WAN environment. In fact, perhaps the most famous WAN of all, the Internet, is a partial mesh configuration network. Again, in this diagram, I’ve illustrated the partial mesh. As previously, since it’s normally a WAN technology, I’ve used cities to illustrate the nodes.
As you can see, not all the cities are connected to all the other cities. Instead, couple of the cities are only connected to two other cities, while a couple of them are connected to three other cities. The result is that we do not have to have a full mesh where every city is connected to every other city, thereby reducing the cost of this particular configuration.
The bus topology is the next one I want to talk about. The bus topology is the oldest network technology available.
With the bus technology, all the nodes or computers are connected directly to a main cable that runs through the network called the bus. The bus topology is very simple to put together because you just have a single backbone cable. Then you have each one of the devices on the network connecting to that cable. One of the disadvantages of the bus topology is that only one node or one computer can send a signal at a time. To determine which computer gets to send a signal, the bus topology uses something called contention.
Basically, what happens is each computer on the bus network competes with each other to send the signal first. Each device listens to the network. If it doesn’t hear a signal on a network at that time, it’ll then send out its own signal. Unfortunately, if two devices listen to the network and don’t hear a signal at the same time, they will send their signal at the same time, and this will result in a collision.
A collision is where the two signals hit each other and effectively cancel each other out. The more nodes and the more computers you have on a bus network. The more likely you are to have collisions. After a collision takes place, the contention to see which device gets to send its signal next starts all over again. If you have too many collisions on a network, it can actually bring down the entire network. In other words, if you have sufficient number of collisions on the network, which is sometimes referred to as a network storm. You can actually bring down the entire network just from collisions.
In fact, this is the idea behind many of the denial-of-service attacks you hear about in network security. One thing about bus topologies to remember is that it’s generally recommended for networks with less than 30 nodes. Otherwise, the contentions became so prevalent that you run the risk of bringing down your network. Now, some advantages to bus topology is the least expensive of the various technologies we’re going to look at.
However, on the down side, a single bad node or cable on a bus topology network can bring down the entire network. Perhaps the biggest disadvantage of the bus topology, though, is that it’s not part of the current TIA/EIA 568-C standard for LAN networks.
This diagram illustrates what a bus topology would look like. As you can see, we have the main wire going through the network, and each one of the computers or nodes then connects directly to that main wire, resulting in a bus topology.
When using a bus topology, you need to make sure that there’s a terminator, also called a terminating resistor, on both ends of the main bus cable there. The next topology I’d like to look at is the ring topology. The ring topology is also one of the older network topologies available to us, and it is similar to the bus topology in that you have a single backbone cable that all the nodes connect to or all the computers connect to. The difference is, instead of having that backbone cable just laying out there like in the bus topology, that backbone cable is basically connected to itself to form a ring or a circle.
Packets are then able to move around the circle in a ring pattern. This is where the ring in the ring topology comes from. As the packet moves through the ring, it stops at each node and gives each node the opportunity to send a signal. This results in no contention between the nodes because every node gets a turn to send a signal. If it doesn’t need to send a signal, it will say, “No, I don’t need it,” and pass it on.
If it does need to send a signal, it gets its opportunity to send a signal. While heavy traffic will not bring down a ring topology, heavy traffic can slow it down because if you have multiple computers trying to all send a packet or signal at one time. Then basically each computer has to take its turn. The result is that things can come very slow as the token or the packet is moving around and giving each computer on the ring an opportunity to send its packet.
Also, like the bus topology, a single damaged node or cable can bring down the entire network. Again, the ring topology is not part of the current TIA/EIA 568-C standard for LAN networks. In this diagram, we can see what a ring network looks like. As you can see, our main wire there in the middle is formed in a ring, and each one of our nodes or each one of our computers are then tied into that central ring shaped backbone wire.
As the token goes around the network, it basically gives each computer an opportunity to send a signal. And once it sends its signal, then the next computer gets a chance to send its signal and so on and so forth around the ring.
Latest posts by Hamza Arif (see all)
- Wide area networks (WAN) & Metropolitan area networks (MAN) - August 18, 2018
- Understanding the Internet, intranets, and extranets - August 17, 2018
- Exploring network collisions, CSMA/CD, and CSMA/CA - August 16, 2018